RepoMirrors/haproxy
1
0
Fork 0
mirror of http://git.haproxy.org/git/haproxy.git/ synced 2025-01-20 04:30:46 +00:00
Code Issues Packages Projects Releases Wiki Activity
8f34320e15
haproxy/addons/promex/service-prometheus.c

2215 lines
84 KiB
C
Raw Normal View History

CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
/*
* Promex is a Prometheus exporter for HAProxy
*
* It is highly inspired by the official Prometheus exporter.
* See: https://github.com/prometheus/haproxy_exporter
*
* Copyright 2019 Christopher Faulet <cfaulet@haproxy.com>
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version
* 2 of the License, or (at your option) any later version.
*
*/
REORG: include: move action.h to haproxy/action{,-t}.h List.h was missing for LIST_ADDQ(). A few unneeded includes of action.h were removed from certain files. This one still relies on applet.h and stick-table.h.
2020-06-04 08:15:32 +00:00
#include <haproxy/action-t.h>
CLEANUP: include: tree-wide alphabetical sort of include files This patch fixes all the leftovers from the include cleanup campaign. There were not that many (~400 entries in ~150 files) but it was definitely worth doing it as it revealed a few duplicates.
2020-06-09 07:07:15 +00:00
#include <haproxy/api.h>
REORG: move applet.h to haproxy/applet{,-t}.h The type file was slightly tidied. The cli-specific APPCTX_CLI_ST1_* flag definitions were moved to cli.h. The type file was adjusted to include buf-t.h and not the huge buf.h. A few call places were fixed because they did not need this include.
2020-06-04 17:42:41 +00:00
#include <haproxy/applet.h>
REORG: include: move backend.h to haproxy/backend{,-t}.h The files remained mostly unchanged since they were OK. However, half of the users didn't need to include them, and about as many actually needed to have it and used to find functions like srv_currently_usable() through a long chain that broke when moving the file.
2020-06-04 20:50:02 +00:00
#include <haproxy/backend.h>
REORG: include: move cfgparse.h to haproxy/cfgparse.h There's no point splitting the file in two since only cfgparse uses the types defined there. A few call places were updated and cleaned up. All of them were in C files which register keywords. There is nothing left in common/ now so this directory must not be used anymore.
2020-06-04 22:00:29 +00:00
#include <haproxy/cfgparse.h>
MAJOR: contrib/prometheus-exporter: move health check status to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for health check states to labels values. The diff is quite small thanks to the preparation work from Christopher to allow more flexibility in labels, see commit 5a2f938732126f43bbf4cea5482c01552b0d0314 ("MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones") this is a follow up of commit c6464591a365bfcf509b322bdaa4d608c9395d75 ("MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels"). The main goal being to be better aligned with prometheus use cases in terms of queries. More specifically to health checks, Pierre C. mentioned the possible quirks he had to put in place in order to make use of those metrics through prometheus: <aggregator_function> by(proxy, check_status) (count_values by(proxy, instance) ("check_status", haproxy_server_check_status)) I am perfectly aware this introduces a lot more metrics but I don't see how we can improve the usability without it. The main issue remains in the cardinality of the states which are > 20. Prometheus recommends to stay below a cardinality of 10 for a given metric but I consider our case very specific, because highly linked to the level of precision haproxy exposes. Even before this patch I saw several large production setup (a few hundreds of MB in output) which are making use of the scope parameter to simply ignore the server metrics, so that the scrapping can be faster, and memory consumed on client side not too high. So I believe we should eventually continue in that direction and offer more granularity of filtering of the output. That being said it is already possible to filter out the data on prometheus client side. this is related to github issue #1029 Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-01 12:11:51 +00:00
#include <haproxy/check.h>
REORG: include: move frontend.h to haproxy/frontend.h There was no type file for this one, it only contains frontend_accept().
2020-06-04 09:23:07 +00:00
#include <haproxy/frontend.h>
REORG: include: split global.h into haproxy/global{,-t}.h global.h was one of the messiest files, it has accumulated tons of implicit dependencies and declares many globals that make almost all other file include it. It managed to silence a dependency loop between server.h and proxy.h by being well placed to pre-define the required structs, forcing struct proxy and struct server to be forward-declared in a significant number of files. It was split in to, one which is the global struct definition and the few macros and flags, and the rest containing the functions prototypes. The UNIX_MAX_PATH definition was moved to compat.h.
2020-06-04 15:05:57 +00:00
#include <haproxy/global.h>
REORG: include: split common/http.h into haproxy/http{,-t}.h So the enums and structs were placed into http-t.h and the functions into http.h. This revealed that several files were dependeng on http.h but not including it, as it was silently inherited via other files.
2020-06-02 17:11:26 +00:00
#include <haproxy/http.h>
BUILD: tree-wide: add missing http_ana.h from many places At least 6 files make use of s->txn without including http_ana which defines it. They used to get it from other includes.
2021-10-06 16:56:42 +00:00
#include <haproxy/http_ana.h>
REORG: include: move http_htx.h to haproxy/http_htx{,-t}.h A few includes had to be added, namely list-t.h in the type file and types/proxy.h in the proto file. actions.h was including http-htx.h but didn't need it so it was dropped.
2020-06-04 07:08:41 +00:00
#include <haproxy/http_htx.h>
REORG: include: split common/htx.h into haproxy/htx{,-t}.h Most of the file was a large set of HTX elements manipulation functions and few types, so splitting them allowed to further reduce dependencies and shrink the build time. Doing so revealed that a few files (h2.c, mux_pt.c) needed haproxy/buf.h and were previously getting it through htx.h. They were fixed.
2020-06-03 06:44:35 +00:00
#include <haproxy/htx.h>
REORG: include: move the error reporting functions to from log.h to errors.h Most of the files dealing with error reports have to include log.h in order to access ha_alert(), ha_warning() etc. But while these functions don't depend on anything, log.h depends on a lot of stuff because it deals with log-formats and samples. As a result it's impossible not to embark long dependencies when using ha_warning() or qfprintf(). This patch moves these low-level functions to errors.h, which already defines the error codes used at the same places. About half of the users of log.h could be adjusted, sometimes revealing other issues such as missing tools.h. Interestingly the total preprocessed size shrunk by 4%.
2020-06-05 15:27:29 +00:00
#include <haproxy/list.h>
REORG: include: move listener.h to haproxy/listener{,-t}.h stdlib and list were missing from listener.h, otherwise it was OK.
2020-06-04 12:58:24 +00:00
#include <haproxy/listener.h>
REORG: include: move the error reporting functions to from log.h to errors.h Most of the files dealing with error reports have to include log.h in order to access ha_alert(), ha_warning() etc. But while these functions don't depend on anything, log.h depends on a lot of stuff because it deals with log-formats and samples. As a result it's impossible not to embark long dependencies when using ha_warning() or qfprintf(). This patch moves these low-level functions to errors.h, which already defines the error codes used at the same places. About half of the users of log.h could be adjusted, sometimes revealing other issues such as missing tools.h. Interestingly the total preprocessed size shrunk by 4%.
2020-06-05 15:27:29 +00:00
#include <haproxy/log.h>
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
#include <haproxy/pool.h>
REORG: include: move proxy.h to haproxy/proxy{,-t}.h This one is particularly difficult to split because it provides all the functions used to manipulate a proxy state and to retrieve names or IDs for error reporting, and as such, it was included in 73 files (down to 68 after cleanup). It would deserve a small cleanup though the cut points are not obvious at the moment given the number of structs involved in the struct proxy itself.
2020-06-04 20:29:18 +00:00
#include <haproxy/proxy.h>
REORG: include: move sample.h to haproxy/sample{,-t}.h This one is particularly tricky to move because everyone uses it and it depends on a lot of other types. For example it cannot include arg-t.h and must absolutely only rely on forward declarations to avoid dependency loops between vars -> sample_data -> arg. In order to address this one, it would be nice to split the sample_data part out of sample.h.
2020-06-04 13:33:47 +00:00
#include <haproxy/sample.h>
REORG: rename cs_utils.h to sc_strm.h This file contains all the stream-connector functions that are specific to application layers of type stream. So let's name it accordingly so that it's easier to figure what's located there. The alphabetical ordering of include files was preserved.
2022-05-27 07:25:10 +00:00
#include <haproxy/sc_strm.h>
REORG: include: move server.h to haproxy/server{,-t}.h extern struct dict server_name_dict was moved from the type file to the main file. A handful of inlined functions were moved at the bottom of the file. Call places were updated to use server-t.h when relevant, or to simply drop the entry when not needed.
2020-06-04 21:20:13 +00:00
#include <haproxy/server.h>
REORG: include: move stats.h to haproxy/stats{,-t}.h Just some minor reordering, and the usual cleanup of call places for those which didn't need it. We don't include the whole tools.h into stats-t anymore but just tools-t.h.
2020-06-04 17:58:55 +00:00
#include <haproxy/stats.h>
REORG: stconn: rename conn_stream.{c,h} to stconn.{c,h} There's no more reason for keepin the code and definitions in conn_stream, let's move all that to stconn. The alphabetical ordering of include files was adjusted.
2022-05-27 07:47:12 +00:00
#include <haproxy/stconn.h>
REORG: include: move stream.h to haproxy/stream{,-t}.h This one was not easy because it was embarking many includes with it, which other files would automatically find. At least global.h, arg.h and tools.h were identified. 93 total locations were identified, 8 additional includes had to be added. In the rare files where it was possible to finalize the sorting of includes by adjusting only one or two extra lines, it was done. But all files would need to be rechecked and cleaned up now. It was the last set of files in types/ and proto/ and these directories must not be reused anymore.
2020-06-04 21:46:14 +00:00
#include <haproxy/stream.h>
REORG: include: move task.h to haproxy/task{,-t}.h The TASK_IS_TASKLET() macro was moved to the proto file instead of the type one. The proto part was a bit reordered to remove a number of ugly forward declaration of static inline functions. About a tens of C and H files had their dependency dropped since they were not using anything from task.h.
2020-06-04 15:25:40 +00:00
#include <haproxy/task.h>
BUILD: promex: service-prometheus.c needs tools.h It calls url_decode() and memprintf() but used to inherit them through others.
2021-05-08 10:58:12 +00:00
#include <haproxy/tools.h>
MINOR: contrib/prometheus-exporter: export build_info commit c55a626217a7e676e1cc ("MINOR: contrib/prometheus-exporter: Add missing global and per-server metrics") is renaming two metrics between v2.2 and v2.3: server_idle_connections_current server_idle_connections_limit It is breaking some tools which are making use of those metrics while supporting several haproxy versions. This build_info will permit tools which make use of metrics to be able to match the haproxy version and change the list of expected metrics. This was possible using the haproxy stats socket but not with prometheus export. This patch follows prometheus best pratices to export specific software informations. It is adding a new field `build_info` so we can extend it to other parameters if needed in the future. example output: # HELP haproxy_process_build_info HAProxy build info. # TYPE haproxy_process_build_info gauge haproxy_process_build_info{version="2.4-dev5-2e1a3f-5"} 1 Even though it is not a bugfix, this patch will make more sense when backported up to >= 2.0 Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-08 12:18:06 +00:00
#include <haproxy/version.h>
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
#include <haproxy/xxhash.h>
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MEDIUM: promex: Add a registration mechanism to support modules In this patch we add a registration mechanism for modules. To do so, a module must defined the "promex_module" structure. The dump itself will be based on 2 contexts. One for all the dump and another one for each metric time-series. These contexts are used as restart points when the dump is interrupted. Modules must also implement 6 callback functions: * start_metric_dump(): It is an optional callback function. If defined, it is responsible to initialize the dump context use as the first restart point. * stop_metric_dump(): It is an optional callback function. If defined, it is responsible to deinit the dump context. * metric_info(): This one is mandatory. It returns the info about the metric: name, type and flags and descrition. * start_ts(): This one is mandatory, it initializes the context for a time series for a given metric. This context is the second restart point. * next_ts(): This one is mandatory. It interates on time series for a given metrics. It is also responsible to handle end of a time series and deinit the context. * fill_ts(): It fills info on the time series for a given metric : the labels and the value. In addition, a module must set its name and declare the number of metrics is exposed.
2024-01-30 06:34:54 +00:00
#include <promex/promex.h>
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
/* Prometheus exporter applet states (appctx->st0) */
enum {
PROMEX_ST_INIT = 0, /* initialized */
PROMEX_ST_HEAD, /* send headers before dump */
PROMEX_ST_DUMP, /* dumping stats */
PROMEX_ST_DONE, /* finished */
BUG/MINOR: contrib/prometheus-exporter: Fix applet accordingly to recent changes Since the flag EOI was added on channels, some hidden bugs in the prometheus exporter now leads to error. the visible effect is that responses are truncated. So first of all, channel_add_input() must be called when the response headers and the EOM block are added. To be sure to correctly update the response channel (especially to_forward value). Then the request must really be fully consumed. And finally, the return clause in the switch has been replaced by a break. It was totally wrong to skip the end of the function in the states PROMEX_DONE and PROMEX_ERROR. (Note that PROMEX_ERROR was never used, so it was replaced by PROMEX_END just to ease reading the code). No need to backport this patch, the Prometheus exporter does not exist in early versions.
2019-03-27 14:48:53 +00:00
PROMEX_ST_END, /* treatment terminated */
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
};
/* Prometheus exporter dumper states (appctx->st1) */
enum {
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
PROMEX_DUMPER_INIT = 0, /* initialized */
PROMEX_DUMPER_GLOBAL, /* dump metrics of globals */
PROMEX_DUMPER_FRONT, /* dump metrics of frontend proxies */
PROMEX_DUMPER_BACK, /* dump metrics of backend proxies */
PROMEX_DUMPER_LI, /* dump metrics of listeners */
PROMEX_DUMPER_SRV, /* dump metrics of servers */
MEDIUM: promex/stick-table: Dump stick-table metrics via a promex module Create a promex module to dump stick-table metrics. Thanks to this patch, all references to stick tables were removed from the promex service.
2024-01-31 09:33:55 +00:00
PROMEX_DUMPER_MODULES, /* dump metrics of modules */
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
PROMEX_DUMPER_DONE, /* finished */
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
};
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
struct promex_module_ref {
struct promex_module *mod;
struct list list;
};
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
/* An entry in a headers map */
struct promex_metric_filter {
int exclude;
struct eb32_node node;
};
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
/* the context of the applet */
struct promex_ctx {
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
void *p[4]; /* generic pointers used to save context */
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
unsigned int flags; /* PROMEX_FL_* */
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
unsigned field_num; /* current field number (ST_I_PX_* etc) */
unsigned mod_field_num; /* first field number of the current module (ST_I_PX_* etc) */
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
int obj_state; /* current state among PROMEX_{FRONT|BACK|SRV|LI}_STATE_* */
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
struct list modules; /* list of promex modules to export */
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
struct eb_root filters; /* list of filters to apply on metrics name */
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
};
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
/* The max length for metrics name. It is a hard limit but it should be
CLEANUP: assorted typo fixes in the code and comments This is fifth iteration of typo fixes
2020-03-23 17:28:40 +00:00
* enough.
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
*/
#define PROMEX_MAX_NAME_LEN 128
/* The expected max length for a metric dump, including its header lines. It is
* just a soft limit to avoid extra work. We don't try to dump a metric if less
* than this size is available in the HTX.
*/
#define PROMEX_MAX_METRIC_LENGTH 512
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
/* Global metrics */
MINOR: stats: rename info stats Info are used to expose haproxy global metrics. It is similar to proxy statistics and any other module. As such, rename info indexes using SI_I_INF_* prefix. Also info variable is renamed stat_line_info. Thanks to this, naming is now consistent between info and other statistics. It will help to integrate it as a "global" statistics module.
2024-04-22 07:41:15 +00:00
const struct promex_metric promex_global_metrics[ST_I_INF_MAX] = {
//[ST_I_INF_NAME] ignored
//[ST_I_INF_VERSION], ignored
//[ST_I_INF_RELEASE_DATE] ignored
[ST_I_INF_NBTHREAD] = { .n = IST("nbthread"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_NBPROC] = { .n = IST("nbproc"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_PROCESS_NUM] = { .n = IST("relative_process_id"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
//[ST_I_INF_PID] ignored
//[ST_I_INF_UPTIME] ignored
[ST_I_INF_UPTIME_SEC] = { .n = IST("uptime_seconds"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_START_TIME_SEC] = { .n = IST("start_time_seconds"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
//[ST_I_INF_MEMMAX_MB] ignored
[ST_I_INF_MEMMAX_BYTES] = { .n = IST("max_memory_bytes"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
//[ST_I_INF_POOL_ALLOC_MB] ignored
[ST_I_INF_POOL_ALLOC_BYTES] = { .n = IST("pool_allocated_bytes"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
//[ST_I_INF_POOL_USED_MB] ignored
[ST_I_INF_POOL_USED_BYTES] = { .n = IST("pool_used_bytes"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_POOL_FAILED] = { .n = IST("pool_failures_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_ULIMIT_N] = { .n = IST("max_fds"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_MAXSOCK] = { .n = IST("max_sockets"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_MAXCONN] = { .n = IST("max_connections"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_HARD_MAXCONN] = { .n = IST("hard_max_connections"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_CURR_CONN] = { .n = IST("current_connections"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_CUM_CONN] = { .n = IST("connections_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_CUM_REQ] = { .n = IST("requests_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_MAX_SSL_CONNS] = { .n = IST("max_ssl_connections"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_CURR_SSL_CONNS] = { .n = IST("current_ssl_connections"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_CUM_SSL_CONNS] = { .n = IST("ssl_connections_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_MAXPIPES] = { .n = IST("max_pipes"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_PIPES_USED] = { .n = IST("pipes_used_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_PIPES_FREE] = { .n = IST("pipes_free_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_CONN_RATE] = { .n = IST("current_connection_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_CONN_RATE_LIMIT] = { .n = IST("limit_connection_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_MAX_CONN_RATE] = { .n = IST("max_connection_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SESS_RATE] = { .n = IST("current_session_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SESS_RATE_LIMIT] = { .n = IST("limit_session_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_MAX_SESS_RATE] = { .n = IST("max_session_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SSL_RATE] = { .n = IST("current_ssl_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SSL_RATE_LIMIT] = { .n = IST("limit_ssl_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_MAX_SSL_RATE] = { .n = IST("max_ssl_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SSL_FRONTEND_KEY_RATE] = { .n = IST("current_frontend_ssl_key_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SSL_FRONTEND_MAX_KEY_RATE] = { .n = IST("max_frontend_ssl_key_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SSL_FRONTEND_SESSION_REUSE_PCT] = { .n = IST("frontend_ssl_reuse"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SSL_BACKEND_KEY_RATE] = { .n = IST("current_backend_ssl_key_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SSL_BACKEND_MAX_KEY_RATE] = { .n = IST("max_backend_ssl_key_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SSL_CACHE_LOOKUPS] = { .n = IST("ssl_cache_lookups_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_SSL_CACHE_MISSES] = { .n = IST("ssl_cache_misses_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_COMPRESS_BPS_IN] = { .n = IST("http_comp_bytes_in_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_COMPRESS_BPS_OUT] = { .n = IST("http_comp_bytes_out_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_COMPRESS_BPS_RATE_LIM] = { .n = IST("limit_http_comp"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_ZLIB_MEM_USAGE] = { .n = IST("current_zlib_memory"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_MAX_ZLIB_MEM_USAGE] = { .n = IST("max_zlib_memory"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_TASKS] = { .n = IST("current_tasks"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_RUN_QUEUE] = { .n = IST("current_run_queue"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_IDLE_PCT] = { .n = IST("idle_time_percent"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
//[ST_I_INF_NODE] ignored
//[ST_I_INF_DESCRIPTION] ignored
[ST_I_INF_STOPPING] = { .n = IST("stopping"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_JOBS] = { .n = IST("jobs"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_UNSTOPPABLE_JOBS] = { .n = IST("unstoppable_jobs"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_LISTENERS] = { .n = IST("listeners"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_ACTIVE_PEERS] = { .n = IST("active_peers"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_CONNECTED_PEERS] = { .n = IST("connected_peers"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_DROPPED_LOGS] = { .n = IST("dropped_logs_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_BUSY_POLLING] = { .n = IST("busy_polling_enabled"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_FAILED_RESOLUTIONS] = { .n = IST("failed_resolutions"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_TOTAL_BYTES_OUT] = { .n = IST("bytes_out_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_TOTAL_SPLICED_BYTES_OUT] = { .n = IST("spliced_bytes_out_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_BYTES_OUT_RATE] = { .n = IST("bytes_out_rate"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
//[ST_I_INF_DEBUG_COMMANDS_ISSUED] ignored
[ST_I_INF_CUM_LOG_MSGS] = { .n = IST("recv_logs_total"), .type = PROMEX_MT_COUNTER, .flags = PROMEX_FL_INFO_METRIC },
[ST_I_INF_BUILD_INFO] = { .n = IST("build_info"), .type = PROMEX_MT_GAUGE, .flags = PROMEX_FL_INFO_METRIC },
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
};
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
/* frontend/backend/server fields */
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
const struct promex_metric promex_st_metrics[ST_I_PX_MAX] = {
//[ST_I_PX_PXNAME] ignored
//[ST_I_PX_SVNAME] ignored
[ST_I_PX_QCUR] = { .n = IST("current_queue"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_QMAX] = { .n = IST("max_queue"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_SCUR] = { .n = IST("current_sessions"), .type = PROMEX_MT_GAUGE, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_SMAX] = { .n = IST("max_sessions"), .type = PROMEX_MT_GAUGE, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_SLIM] = { .n = IST("limit_sessions"), .type = PROMEX_MT_GAUGE, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_STOT] = { .n = IST("sessions_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_BIN] = { .n = IST("bytes_in_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_BOUT] = { .n = IST("bytes_out_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_DREQ] = { .n = IST("requests_denied_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_DRESP] = { .n = IST("responses_denied_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_EREQ] = { .n = IST("request_errors_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC ) },
[ST_I_PX_ECON] = { .n = IST("connection_errors_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_ERESP] = { .n = IST("response_errors_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_WRETR] = { .n = IST("retry_warnings_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_WREDIS] = { .n = IST("redispatch_warnings_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_STATUS] = { .n = IST("status"), .type = PROMEX_MT_GAUGE, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_WEIGHT] = { .n = IST("weight"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_ACT] = { .n = IST("active_servers"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_BCK] = { .n = IST("backup_servers"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_CHKFAIL] = { .n = IST("check_failures_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_CHKDOWN] = { .n = IST("check_up_down_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_LASTCHG] = { .n = IST("check_last_change_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_DOWNTIME] = { .n = IST("downtime_seconds_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_QLIMIT] = { .n = IST("queue_limit"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
//[ST_I_PX_PID] ignored
//[ST_I_PX_IID] ignored
//[ST_I_PX_SID] ignored
[ST_I_PX_THROTTLE] = { .n = IST("current_throttle"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_LBTOT] = { .n = IST("loadbalanced_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
//[ST_I_PX_TRACKED] ignored
//[ST_I_PX_TYPE] ignored
//[ST_I_PX_RATE] ignored
[ST_I_PX_RATE_LIM] = { .n = IST("limit_session_rate"), .type = PROMEX_MT_GAUGE, .flags = (PROMEX_FL_FRONT_METRIC ) },
[ST_I_PX_RATE_MAX] = { .n = IST("max_session_rate"), .type = PROMEX_MT_GAUGE, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_CHECK_STATUS] = { .n = IST("check_status"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_CHECK_CODE] = { .n = IST("check_code"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_CHECK_DURATION] = { .n = IST("check_duration_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_HRSP_1XX] = { .n = IST("http_responses_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_HRSP_2XX] = { .n = IST("http_responses_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_HRSP_3XX] = { .n = IST("http_responses_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_HRSP_4XX] = { .n = IST("http_responses_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_HRSP_5XX] = { .n = IST("http_responses_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_HRSP_OTHER] = { .n = IST("http_responses_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
//[ST_I_PX_HANAFAIL] ignored
//[ST_I_PX_REQ_RATE] ignored
[ST_I_PX_REQ_RATE_MAX] = { .n = IST("http_requests_rate_max"), .type = PROMEX_MT_GAUGE, .flags = (PROMEX_FL_FRONT_METRIC ) },
[ST_I_PX_REQ_TOT] = { .n = IST("http_requests_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_CLI_ABRT] = { .n = IST("client_aborts_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_SRV_ABRT] = { .n = IST("server_aborts_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_COMP_IN] = { .n = IST("http_comp_bytes_in_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_COMP_OUT] = { .n = IST("http_comp_bytes_out_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_COMP_BYP] = { .n = IST("http_comp_bytes_bypassed_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_COMP_RSP] = { .n = IST("http_comp_responses_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_LASTSESS] = { .n = IST("last_session_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
//[ST_I_PX_LAST_CHK] ignored
//[ST_I_PX_LAST_AGT] ignored
[ST_I_PX_QTIME] = { .n = IST("queue_time_average_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_CTIME] = { .n = IST("connect_time_average_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_RTIME] = { .n = IST("response_time_average_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_TTIME] = { .n = IST("total_time_average_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
//[ST_I_PX_AGENT_STATUS] ignored
//[ST_I_PX_AGENT_CODE] ignored
//[ST_I_PX_AGENT_DURATION] ignored
//[ST_I_PX_CHECK_DESC] ignored
//[ST_I_PX_AGENT_DESC] ignored
//[ST_I_PX_CHECK_RISE] ignored
//[ST_I_PX_CHECK_FALL] ignored
//[ST_I_PX_CHECK_HEALTH] ignored
//[ST_I_PX_AGENT_RISE] ignored
//[ST_I_PX_AGENT_FALL] ignored
//[ST_I_PX_AGENT_HEALTH] ignored
//[ST_I_PX_ADDR] ignored
//[ST_I_PX_COOKIE] ignored
//[ST_I_PX_MODE] ignored
//[ST_I_PX_ALGO] ignored
//[ST_I_PX_CONN_RATE] ignored
[ST_I_PX_CONN_RATE_MAX] = { .n = IST("connections_rate_max"), .type = PROMEX_MT_GAUGE, .flags = (PROMEX_FL_FRONT_METRIC ) },
[ST_I_PX_CONN_TOT] = { .n = IST("connections_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC ) },
[ST_I_PX_INTERCEPTED] = { .n = IST("intercepted_requests_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC ) },
[ST_I_PX_DCON] = { .n = IST("denied_connections_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC ) },
[ST_I_PX_DSES] = { .n = IST("denied_sessions_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC ) },
[ST_I_PX_WREW] = { .n = IST("failed_header_rewriting_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_CONNECT] = { .n = IST("connection_attempts_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_REUSE] = { .n = IST("connection_reuses_total"), .type = PROMEX_MT_COUNTER, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_CACHE_LOOKUPS] = { .n = IST("http_cache_lookups_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_CACHE_HITS] = { .n = IST("http_cache_hits_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_SRV_ICUR] = { .n = IST("idle_connections_current"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_SRV_ILIM] = { .n = IST("idle_connections_limit"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_QT_MAX] = { .n = IST("max_queue_time_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_CT_MAX] = { .n = IST("max_connect_time_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_RT_MAX] = { .n = IST("max_response_time_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_TT_MAX] = { .n = IST("max_total_time_seconds"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_EINT] = { .n = IST("internal_errors_total"), .type = PROMEX_MT_COUNTER, .flags = (PROMEX_FL_FRONT_METRIC | PROMEX_FL_LI_METRIC | PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_IDLE_CONN_CUR] = { .n = IST("unsafe_idle_connections_current"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_SAFE_CONN_CUR] = { .n = IST("safe_idle_connections_current"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_USED_CONN_CUR] = { .n = IST("used_connections_current"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_NEED_CONN_EST] = { .n = IST("need_connections_current"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_SRV_METRIC) },
[ST_I_PX_UWEIGHT] = { .n = IST("uweight"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC | PROMEX_FL_SRV_METRIC) },
[ST_I_PX_AGG_SRV_CHECK_STATUS] = { .n = IST("agg_server_check_status"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_AGG_SRV_STATUS ] = { .n = IST("agg_server_status"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC ) },
[ST_I_PX_AGG_CHECK_STATUS] = { .n = IST("agg_check_status"), .type = PROMEX_MT_GAUGE, .flags = ( PROMEX_FL_BACK_METRIC ) },
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
};
MINOR: promex: Add support for specialized front/back/li/srv metric names Depending on the scope, metrics can have different names. For instance, the number of active/backend servers are reported with "haproxy_backend_active_servers"/"haproxy_backend_backup_servers" metric names in the backend scope while it should be "haproxy_server_active"/"haproxy_server_backup" in the server scope. To be able to support different names depending on the scope for the same metric, arrays of ISTs were added, one by scope (front, back, listen, server). These arrays only contain names overriding the default ones. Note: the exemple above is not supported for now and is the reason for this commit.
2023-12-04 08:00:18 +00:00
/* Specialized frontend metric names, to override default ones */
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
const struct ist promex_st_front_metrics_names[ST_I_PX_MAX] = {
MINOR: promex: Add support for specialized front/back/li/srv metric names Depending on the scope, metrics can have different names. For instance, the number of active/backend servers are reported with "haproxy_backend_active_servers"/"haproxy_backend_backup_servers" metric names in the backend scope while it should be "haproxy_server_active"/"haproxy_server_backup" in the server scope. To be able to support different names depending on the scope for the same metric, arrays of ISTs were added, one by scope (front, back, listen, server). These arrays only contain names overriding the default ones. Note: the exemple above is not supported for now and is the reason for this commit.
2023-12-04 08:00:18 +00:00
};
CLEANUP: assorted typo fixes in the code and comments This is 38th iteration of typo fixes
2023-12-29 19:05:20 +00:00
/* Specialized backend metric names, to override default ones */
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
const struct ist promex_st_back_metrics_names[ST_I_PX_MAX] = {
MINOR: promex: Add support for specialized front/back/li/srv metric names Depending on the scope, metrics can have different names. For instance, the number of active/backend servers are reported with "haproxy_backend_active_servers"/"haproxy_backend_backup_servers" metric names in the backend scope while it should be "haproxy_server_active"/"haproxy_server_backup" in the server scope. To be able to support different names depending on the scope for the same metric, arrays of ISTs were added, one by scope (front, back, listen, server). These arrays only contain names overriding the default ones. Note: the exemple above is not supported for now and is the reason for this commit.
2023-12-04 08:00:18 +00:00
};
/* Specialized listener metric names, to override default ones */
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
const struct ist promex_st_li_metrics_names[ST_I_PX_MAX] = {
MINOR: promex: Add support for specialized front/back/li/srv metric names Depending on the scope, metrics can have different names. For instance, the number of active/backend servers are reported with "haproxy_backend_active_servers"/"haproxy_backend_backup_servers" metric names in the backend scope while it should be "haproxy_server_active"/"haproxy_server_backup" in the server scope. To be able to support different names depending on the scope for the same metric, arrays of ISTs were added, one by scope (front, back, listen, server). These arrays only contain names overriding the default ones. Note: the exemple above is not supported for now and is the reason for this commit.
2023-12-04 08:00:18 +00:00
};
/* Specialized server metric names, to override default ones */
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
const struct ist promex_st_srv_metrics_names[ST_I_PX_MAX] = {
[ST_I_PX_ACT] = IST("active"),
[ST_I_PX_BCK] = IST("backup"),
MINOR: promex: Add support for specialized front/back/li/srv metric names Depending on the scope, metrics can have different names. For instance, the number of active/backend servers are reported with "haproxy_backend_active_servers"/"haproxy_backend_backup_servers" metric names in the backend scope while it should be "haproxy_server_active"/"haproxy_server_backup" in the server scope. To be able to support different names depending on the scope for the same metric, arrays of ISTs were added, one by scope (front, back, listen, server). These arrays only contain names overriding the default ones. Note: the exemple above is not supported for now and is the reason for this commit.
2023-12-04 08:00:18 +00:00
};
CLEANUP: assorted typo fixes in the code and comments This is 17th iteration of typo fixes
2021-02-06 17:29:08 +00:00
/* Description of overridden stats fields */
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
const struct ist promex_st_metric_desc[ST_I_PX_MAX] = {
[ST_I_PX_STATUS] = IST("Current status of the service, per state label value."),
[ST_I_PX_CHECK_STATUS] = IST("Status of last health check, per state label value."),
[ST_I_PX_CHECK_CODE] = IST("layer5-7 code, if available of the last health check."),
[ST_I_PX_CHECK_DURATION] = IST("Total duration of the latest server health check, in seconds."),
[ST_I_PX_QTIME] = IST("Avg. queue time for last 1024 successful connections."),
[ST_I_PX_CTIME] = IST("Avg. connect time for last 1024 successful connections."),
[ST_I_PX_RTIME] = IST("Avg. response time for last 1024 successful connections."),
[ST_I_PX_TTIME] = IST("Avg. total time for last 1024 successful connections."),
[ST_I_PX_QT_MAX] = IST("Maximum observed time spent in the queue"),
[ST_I_PX_CT_MAX] = IST("Maximum observed time spent waiting for a connection to complete"),
[ST_I_PX_RT_MAX] = IST("Maximum observed time spent waiting for a server response"),
[ST_I_PX_TT_MAX] = IST("Maximum observed total request+response time (request+queue+connect+response+processing)"),
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
};
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
/* Specific labels for all ST_I_PX_HRSP_* fields */
const struct ist promex_hrsp_code[1 + ST_I_PX_HRSP_OTHER - ST_I_PX_HRSP_1XX] = {
[ST_I_PX_HRSP_1XX - ST_I_PX_HRSP_1XX] = IST("1xx"),
[ST_I_PX_HRSP_2XX - ST_I_PX_HRSP_1XX] = IST("2xx"),
[ST_I_PX_HRSP_3XX - ST_I_PX_HRSP_1XX] = IST("3xx"),
[ST_I_PX_HRSP_4XX - ST_I_PX_HRSP_1XX] = IST("4xx"),
[ST_I_PX_HRSP_5XX - ST_I_PX_HRSP_1XX] = IST("5xx"),
[ST_I_PX_HRSP_OTHER - ST_I_PX_HRSP_1XX] = IST("other"),
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
};
MINOR: contrib/prometheus-exporter: declare states for objects in preparation to change state gauge values as labels, declare them as enum associated with the string definition Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:16 +00:00
enum promex_front_state {
PROMEX_FRONT_STATE_DOWN = 0,
PROMEX_FRONT_STATE_UP,
PROMEX_FRONT_STATE_COUNT /* must be last */
};
const struct ist promex_front_st[PROMEX_FRONT_STATE_COUNT] = {
[PROMEX_FRONT_STATE_DOWN] = IST("DOWN"),
[PROMEX_FRONT_STATE_UP] = IST("UP"),
};
enum promex_back_state {
PROMEX_BACK_STATE_DOWN = 0,
PROMEX_BACK_STATE_UP,
PROMEX_BACK_STATE_COUNT /* must be last */
};
const struct ist promex_back_st[PROMEX_BACK_STATE_COUNT] = {
[PROMEX_BACK_STATE_DOWN] = IST("DOWN"),
[PROMEX_BACK_STATE_UP] = IST("UP"),
};
enum promex_srv_state {
PROMEX_SRV_STATE_DOWN = 0,
PROMEX_SRV_STATE_UP,
PROMEX_SRV_STATE_MAINT,
PROMEX_SRV_STATE_DRAIN,
PROMEX_SRV_STATE_NOLB,
PROMEX_SRV_STATE_COUNT /* must be last */
};
const struct ist promex_srv_st[PROMEX_SRV_STATE_COUNT] = {
[PROMEX_SRV_STATE_DOWN] = IST("DOWN"),
[PROMEX_SRV_STATE_UP] = IST("UP"),
[PROMEX_SRV_STATE_MAINT] = IST("MAINT"),
[PROMEX_SRV_STATE_DRAIN] = IST("DRAIN"),
[PROMEX_SRV_STATE_NOLB] = IST("NOLB"),
};
MEDIUM: promex: Add a registration mechanism to support modules In this patch we add a registration mechanism for modules. To do so, a module must defined the "promex_module" structure. The dump itself will be based on 2 contexts. One for all the dump and another one for each metric time-series. These contexts are used as restart points when the dump is interrupted. Modules must also implement 6 callback functions: * start_metric_dump(): It is an optional callback function. If defined, it is responsible to initialize the dump context use as the first restart point. * stop_metric_dump(): It is an optional callback function. If defined, it is responsible to deinit the dump context. * metric_info(): This one is mandatory. It returns the info about the metric: name, type and flags and descrition. * start_ts(): This one is mandatory, it initializes the context for a time series for a given metric. This context is the second restart point. * next_ts(): This one is mandatory. It interates on time series for a given metrics. It is also responsible to handle end of a time series and deinit the context. * fill_ts(): It fills info on the time series for a given metric : the labels and the value. In addition, a module must set its name and declare the number of metrics is exposed.
2024-01-30 06:34:54 +00:00
struct list promex_module_list = LIST_HEAD_INIT(promex_module_list);
void promex_register_module(struct promex_module *m)
{
LIST_APPEND(&promex_module_list, &m->list);
}
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
/* Pools used to allocate ref on Promex modules and filters */
DECLARE_STATIC_POOL(pool_head_promex_mod_ref, "promex_module_ref", sizeof(struct promex_module_ref));
DECLARE_STATIC_POOL(pool_head_promex_metric_flt, "promex_metric_filter", sizeof(struct promex_metric_filter));
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
MINOR: contrib/prometheus-exporter: declare states for objects in preparation to change state gauge values as labels, declare them as enum associated with the string definition Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:16 +00:00
/* Return the server status. */
enum promex_srv_state promex_srv_status(struct server *sv)
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
{
MINOR: contrib/prometheus-exporter: declare states for objects in preparation to change state gauge values as labels, declare them as enum associated with the string definition Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:16 +00:00
int state = PROMEX_SRV_STATE_DOWN;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
if (sv->cur_state == SRV_ST_RUNNING || sv->cur_state == SRV_ST_STARTING) {
MINOR: contrib/prometheus-exporter: declare states for objects in preparation to change state gauge values as labels, declare them as enum associated with the string definition Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:16 +00:00
state = PROMEX_SRV_STATE_UP;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
if (sv->cur_admin & SRV_ADMF_DRAIN)
MINOR: contrib/prometheus-exporter: declare states for objects in preparation to change state gauge values as labels, declare them as enum associated with the string definition Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:16 +00:00
state = PROMEX_SRV_STATE_DRAIN;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MINOR: contrib/prometheus-exporter: Report DRAIN/MAINT/NOLB status for servers Now, following status are reported for servers:0=DOWN, 1=UP, 2=MAINT, 3=DRAIN, 4=NOLB. It is linked to the github issue #255. Thanks to Mickaël Martin. If needed, this patch may be backported to 2.0.
2019-09-06 14:10:19 +00:00
else if (sv->cur_state == SRV_ST_STOPPING)
MINOR: contrib/prometheus-exporter: declare states for objects in preparation to change state gauge values as labels, declare them as enum associated with the string definition Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:16 +00:00
state = PROMEX_SRV_STATE_NOLB;
MINOR: contrib/prometheus-exporter: Report DRAIN/MAINT/NOLB status for servers Now, following status are reported for servers:0=DOWN, 1=UP, 2=MAINT, 3=DRAIN, 4=NOLB. It is linked to the github issue #255. Thanks to Mickaël Martin. If needed, this patch may be backported to 2.0.
2019-09-06 14:10:19 +00:00
if (sv->cur_admin & SRV_ADMF_MAINT)
MINOR: contrib/prometheus-exporter: declare states for objects in preparation to change state gauge values as labels, declare them as enum associated with the string definition Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:16 +00:00
state = PROMEX_SRV_STATE_MAINT;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return state;
}
/* Convert a field to its string representation and write it in <out>, followed
* by a newline, if there is enough space. non-numeric value are converted in
MINOR: contrib/prometheus-exporter: better output of Not-a-Number Not necessarily mandatory but I saw a few prometheus client parsing only `NaN`. Also most librarries do output `NaN` Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-22 20:09:47 +00:00
* "NaN" because Prometheus only support numerical values (but it is unexepceted
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
* to process this kind of value). It returns 1 on success. Otherwise, it
* returns 0. The buffer's length must not exceed <max> value.
*/
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
static int promex_ts_val_to_str(struct buffer *out, struct field *f, size_t max)
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
{
int ret = 0;
switch (field_format(f, 0)) {
MINOR: contrib/prometheus-exporter: better output of Not-a-Number Not necessarily mandatory but I saw a few prometheus client parsing only `NaN`. Also most librarries do output `NaN` Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-22 20:09:47 +00:00
case FF_EMPTY: ret = chunk_strcat(out, "NaN\n"); break;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
case FF_S32: ret = chunk_appendf(out, "%d\n", f->u.s32); break;
case FF_U32: ret = chunk_appendf(out, "%u\n", f->u.u32); break;
case FF_S64: ret = chunk_appendf(out, "%lld\n", (long long)f->u.s64); break;
case FF_U64: ret = chunk_appendf(out, "%llu\n", (unsigned long long)f->u.u64); break;
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
case FF_FLT: ret = chunk_appendf(out, "%f\n", f->u.flt); break;
MINOR: contrib/prometheus-exporter: better output of Not-a-Number Not necessarily mandatory but I saw a few prometheus client parsing only `NaN`. Also most librarries do output `NaN` Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-22 20:09:47 +00:00
case FF_STR: ret = chunk_strcat(out, "NaN\n"); break;
default: ret = chunk_strcat(out, "NaN\n"); break;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
if (!ret || out->data > max)
return 0;
return 1;
}
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
/* Dump the time series header lines for the metric <name>. It is its #HELP and #TYPE
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
* strings. It returns 1 on success. Otherwise, if <out> length exceeds <max>,
* it returns 0.
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
*/
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
static int promex_dump_ts_header(const struct ist name, const struct ist desc, enum promex_mt_type type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
struct ist *out, size_t max)
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
{
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
struct ist t;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
switch (type) {
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
case PROMEX_MT_COUNTER:
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
t = ist("counter");
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
break;
default:
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
t = ist("gauge");
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MINOR: contrib/prometheus-exporter: merge info description from stats Now that units are coherent we can merge the info description from haproxy stats. Description were not always the same, but I guess we may eventually improve them in the future. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-15 21:41:39 +00:00
if (istcat(out, ist("# HELP "), max) == -1 ||
istcat(out, name, max) == -1 ||
istcat(out, ist(" "), max) == -1)
goto full;
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (istcat(out, ist("# HELP "), max) == -1 ||
istcat(out, name, max) == -1 ||
istcat(out, ist(" "), max) == -1 ||
istcat(out, desc, max) == -1)
goto full;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (istcat(out, ist("\n# TYPE "), max) == -1 ||
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
istcat(out, name, max) == -1 ||
istcat(out, ist(" "), max) == -1 ||
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
istcat(out, t, max) == -1 ||
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
istcat(out, ist("\n"), max) == -1)
goto full;
return 1;
full:
return 0;
}
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
/* Dump the time series for the metric <name>. It starts by the metric name followed by
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
* its labels (proxy name, server name...) between braces and finally its
* value. If not already done, the header lines are dumped first. It returns 1
* on success. Otherwise if <out> length exceeds <max>, it returns 0.
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
*/
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
static int promex_dump_ts(struct appctx *appctx, struct ist prefix,
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
const struct ist name, const struct ist desc, enum promex_mt_type type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
struct field *val, struct promex_label *labels, struct ist *out, size_t max)
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
{
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
struct ist n = { .ptr = (char[PROMEX_MAX_NAME_LEN]){ 0 }, .len = 0 };
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
struct promex_ctx *ctx = appctx->svcctx;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
size_t len = out->len;
if (out->len + PROMEX_MAX_METRIC_LENGTH > max)
return 0;
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
/* Fill the metric name */
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
istcat(&n, prefix, PROMEX_MAX_NAME_LEN);
istcat(&n, name, PROMEX_MAX_NAME_LEN);
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
if ((ctx->flags & PROMEX_FL_METRIC_HDR) &&
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
!promex_dump_ts_header(n, desc, type, out, max))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto full;
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (istcat(out, n, max) == -1)
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
goto full;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
if (isttest(labels[0].name)) {
int i;
if (istcat(out, ist("{"), max) == -1)
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto full;
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
MINOR: promex: Always limit the number of labels dumped for each metric It was not an issue since now, be a way to register modules on promex will be added. Thus it is important to add some extra checks. Here, we take care to never dump more than the max labels allowed.
2024-01-29 16:19:28 +00:00
for (i = 0; i < PROMEX_MAX_LABELS && isttest(labels[i].name); i++) {
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
if (!isttest(labels[i].value))
continue;
if ((i && istcat(out, ist(","), max) == -1) ||
istcat(out, labels[i].name, max) == -1 ||
istcat(out, ist("=\""), max) == -1 ||
istcat(out, labels[i].value, max) == -1 ||
istcat(out, ist("\""), max) == -1)
goto full;
}
if (istcat(out, ist("}"), max) == -1)
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto full;
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
if (istcat(out, ist(" "), max) == -1)
goto full;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
trash.data = out->len;
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
if (!promex_ts_val_to_str(&trash, val, max))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto full;
out->len = trash.data;
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~PROMEX_FL_METRIC_HDR;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return 1;
full:
// Restore previous length
out->len = len;
return 0;
}
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
static int promex_filter_metric(struct appctx *appctx, struct ist prefix, struct ist name)
{
struct promex_ctx *ctx = appctx->svcctx;
struct eb32_node *node;
struct promex_metric_filter *flt;
unsigned int hash;
XXH32_state_t state;
if (!eb_is_empty(&ctx->filters)) {
XXH32_reset(&state, 0);
XXH32_update(&state, istptr(prefix), istlen(prefix));
XXH32_update(&state, istptr(name), istlen(name));
hash = XXH32_digest(&state);
node = eb32_lookup(&ctx->filters, hash);
if (node) {
flt = container_of(node, typeof(*flt), node);
if (flt->exclude)
return 1;
}
else if (!(ctx->flags & PROMEX_FL_INC_METRIC_BY_DEFAULT))
return 1;
}
return 0;
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
CLEANUP: assorted typo fixes in the code and comments This is fifth iteration of typo fixes
2020-03-23 17:28:40 +00:00
/* Dump global metrics (prefixed by "haproxy_process_"). It returns 1 on success,
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
* 0 if <htx> is full and -1 in case of any error. */
static int promex_dump_global_metrics(struct appctx *appctx, struct htx *htx)
{
static struct ist prefix = IST("haproxy_process_");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
struct promex_ctx *ctx = appctx->svcctx;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
struct field val;
CLEANUP: applet: rename appctx_cs() to appctx_sc() It returns a stream connector, not a conn_stream anymore, so let's fix its name.
2022-05-27 09:08:15 +00:00
struct channel *chn = sc_ic(appctx_sc(appctx));
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
struct ist name, desc, out = ist2(trash.area, 0);
BUG/MINOR: contrib/prometheus-exporter: Respect the reserve when data are sent The previous commit e6cdfe574 ("BUG/MINOR: contrib/prometheus-exporter: Don't use channel_htx_recv_max()") is buggy. The buffer's reserve must be respected. This patch must be backported to 2.0 and 1.9.
2019-07-03 09:43:17 +00:00
size_t max = htx_get_max_blksz(htx, channel_htx_recv_max(chn, htx));
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
int ret = 1;
MINOR: stats: rename info stats Info are used to expose haproxy global metrics. It is similar to proxy statistics and any other module. As such, rename info indexes using SI_I_INF_* prefix. Also info variable is renamed stat_line_info. Thanks to this, naming is now consistent between info and other statistics. It will help to integrate it as a "global" statistics module.
2024-04-22 07:41:15 +00:00
if (!stats_fill_info(stat_line_info, ST_I_INF_MAX, 0))
MINOR: contrib/prometheus-exporter: use fill_info for process dump use `stats_fill_info` when possible to avoid duplicating code. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-11 19:07:49 +00:00
return -1;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MINOR: stats: rename info stats Info are used to expose haproxy global metrics. It is similar to proxy statistics and any other module. As such, rename info indexes using SI_I_INF_* prefix. Also info variable is renamed stat_line_info. Thanks to this, naming is now consistent between info and other statistics. It will help to integrate it as a "global" statistics module.
2024-04-22 07:41:15 +00:00
for (; ctx->field_num < ST_I_INF_MAX; ctx->field_num++) {
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
struct promex_label labels[PROMEX_MAX_LABELS-1] = {};
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
if (!(promex_global_metrics[ctx->field_num].flags & ctx->flags))
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
continue;
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
name = promex_global_metrics[ctx->field_num].n;
MINOR: stats: update ambiguous "metrics" naming to "stat_cols" The name "metrics" was chosen to represent the various list of haproxy exposed statistics. However, it is deemed as ambiguous as some stats are indeed metric in the true sense, but some are not, as highlighted by various "enum field_origin" values. Replace it by the new name "stat_cols" for statistic columns. Along with the already existing notion of stat lines it should better reflect its purpose.
2024-04-23 09:06:00 +00:00
desc = ist(stat_cols_info[ctx->field_num].desc);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, name))
continue;
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
switch (ctx->field_num) {
MINOR: stats: rename info stats Info are used to expose haproxy global metrics. It is similar to proxy statistics and any other module. As such, rename info indexes using SI_I_INF_* prefix. Also info variable is renamed stat_line_info. Thanks to this, naming is now consistent between info and other statistics. It will help to integrate it as a "global" statistics module.
2024-04-22 07:41:15 +00:00
case ST_I_INF_BUILD_INFO:
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[0].name = ist("version");
labels[0].value = ist(HAPROXY_VERSION);
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_u32(FN_GAUGE, 1);
MINOR: contrib/prometheus-exporter: export build_info commit c55a626217a7e676e1cc ("MINOR: contrib/prometheus-exporter: Add missing global and per-server metrics") is renaming two metrics between v2.2 and v2.3: server_idle_connections_current server_idle_connections_limit It is breaking some tools which are making use of those metrics while supporting several haproxy versions. This build_info will permit tools which make use of metrics to be able to match the haproxy version and change the list of expected metrics. This was possible using the haproxy stats socket but not with prometheus export. This patch follows prometheus best pratices to export specific software informations. It is adding a new field `build_info` so we can extend it to other parameters if needed in the future. example output: # HELP haproxy_process_build_info HAProxy build info. # TYPE haproxy_process_build_info gauge haproxy_process_build_info{version="2.4-dev5-2e1a3f-5"} 1 Even though it is not a bugfix, this patch will make more sense when backported up to >= 2.0 Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-08 12:18:06 +00:00
break;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
default:
MINOR: stats: rename info stats Info are used to expose haproxy global metrics. It is similar to proxy statistics and any other module. As such, rename info indexes using SI_I_INF_* prefix. Also info variable is renamed stat_line_info. Thanks to this, naming is now consistent between info and other statistics. It will help to integrate it as a "global" statistics module.
2024-04-22 07:41:15 +00:00
val = stat_line_info[ctx->field_num];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_global_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto full;
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_METRIC_HDR;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
end:
BUG/MINOR: contrib/prometheus-exporter: Don't try to add empty data blocks When the response buffer is full and nothing more can be inserted, it is important to not try to insert an empty data block. Otherwise, when the function channel_add_input() is called, the flag CF_READ_PARTIAL is set on the response channel while nothing was read and the stream is uselessly woken up. Finally, we have loop while the response buffer is full. This patch must be backported to 2.0.
2019-07-04 08:03:28 +00:00
if (out.len) {
if (!htx_add_data_atonce(htx, out))
return -1; /* Unexpected and unrecoverable error */
channel_add_input(chn, out.len);
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return ret;
full:
ret = 0;
goto end;
}
CLEANUP: assorted typo fixes in the code and comments This is fifth iteration of typo fixes
2020-03-23 17:28:40 +00:00
/* Dump frontends metrics (prefixed by "haproxy_frontend_"). It returns 1 on success,
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
* 0 if <htx> is full and -1 in case of any error. */
static int promex_dump_front_metrics(struct appctx *appctx, struct htx *htx)
{
static struct ist prefix = IST("haproxy_frontend_");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
struct promex_ctx *ctx = appctx->svcctx;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
struct proxy *px = ctx->p[0];
struct stats_module *mod = ctx->p[1];
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
struct field val;
CLEANUP: applet: rename appctx_cs() to appctx_sc() It returns a stream connector, not a conn_stream anymore, so let's fix its name.
2022-05-27 09:08:15 +00:00
struct channel *chn = sc_ic(appctx_sc(appctx));
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
struct ist name, desc, out = ist2(trash.area, 0);
BUG/MINOR: contrib/prometheus-exporter: Respect the reserve when data are sent The previous commit e6cdfe574 ("BUG/MINOR: contrib/prometheus-exporter: Don't use channel_htx_recv_max()") is buggy. The buffer's reserve must be respected. This patch must be backported to 2.0 and 1.9.
2019-07-03 09:43:17 +00:00
size_t max = htx_get_max_blksz(htx, channel_htx_recv_max(chn, htx));
MINOR: stats: rename ambiguous stat_l and stat_count Statistics were extended with the introduction of stats module. This mechanism allows to expose various metrics for several haproxy components. As a consequence of this, some static variables were transformed to dynamic ones to be able to regroup all statistics definition. Rename these variables with more explicit naming : * stat_lines can be used to generate one line of statistics for any module using struct field as value * metrics and metrics_len are used to stored description of metrics indexed by module Note that info is not integrated in the statistics module mechanism. However, it could be done in the future to better reflect its purpose.
2024-04-22 09:19:17 +00:00
struct field *stats = stat_lines[STATS_DOMAIN_PROXY];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
int ret = 1;
MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for frontend/backend/servers states to labels values. the main motivation being I realised it is very difficult to make use of it without hard coded quirks on prometheus client side; especially because the main use is often to group by state, which is harder when the state is the value of the metric. in order to achieve that we iterate on the status metric to generate labels, and so as many metrics. this is the first step to resolve github issue #1029 A second step should address health check states. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:17 +00:00
enum promex_front_state state;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
for (;ctx->field_num < ST_I_PX_MAX; ctx->field_num++) {
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
if (!(promex_st_metrics[ctx->field_num].flags & ctx->flags))
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
continue;
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
name = promex_st_front_metrics_names[ctx->field_num];
desc = promex_st_metric_desc[ctx->field_num];
if (!isttest(name))
name = promex_st_metrics[ctx->field_num].n;
if (!isttest(desc))
MINOR: stats: update ambiguous "metrics" naming to "stat_cols" The name "metrics" was chosen to represent the various list of haproxy exposed statistics. However, it is deemed as ambiguous as some stats are indeed metric in the true sense, but some are not, as highlighted by various "enum field_origin" values. Replace it by the new name "stat_cols" for statistic columns. Along with the already existing notion of stat lines it should better reflect its purpose.
2024-04-23 09:06:00 +00:00
desc = ist(stat_cols_px[ctx->field_num].desc);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, name))
continue;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!px)
px = proxies_list;
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
while (px) {
struct promex_label labels[PROMEX_MAX_LABELS-1] = {};
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[0].name = ist("proxy");
labels[0].value = ist2(px->id, strlen(px->id));
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
/* skip the disabled proxies, global frontend and non-networked ones */
MINOR: proxy: Introduce proxy flags to replace disabled bitfield This change is required to support TCP/HTTP rules in defaults sections. The 'disabled' bitfield in the proxy structure, used to know if a proxy is disabled or stopped, is replaced a generic bitfield named 'flags'. PR_DISABLED and PR_STOPPED flags are renamed to PR_FL_DISABLED and PR_FL_STOPPED respectively. In addition, everywhere there is a test to know if a proxy is disabled or stopped, there is now a bitwise AND operation on PR_FL_DISABLED and/or PR_FL_STOPPED flags.
2021-10-06 12:24:19 +00:00
if ((px->flags & PR_FL_DISABLED) || px->uuid <= 0 || !(px->cap & PR_CAP_FE))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto next_px;
MINOR: stats: define stats-file output format support Prepare stats function to handle a new format labelled "stats-file". Its purpose is to generate a statistics dump with a format closed from the CSV output. Such output will be then used to preload haproxy internal counters on process startup. stats-file output differs from a standard CSV on several points. First, only an excerpt of all statistics is outputted. All values that does not make sense to preload are excluded. For the moment, stats-file only list stats fully defined via "struct stat_col" method. Contrary to a CSV, sll columns of a stats-file will be filled. As such, empty field value is used to mark stats which should not be outputted. Some adaptation specifics to stats-file are necessary into me_generate_field(). First, stats-file will output separatedly values from frontend and backend sides with their own respective set of columns. As such, an empty field value is returned if stat is not defined for either frontend/listener, or backend/server when outputting the other side. Also, as stats-file does not support empty column, stcol_hide() is not used for it. A minor adjustement was necessary for stats_fill_fe_line() to pass context flags. This is necessary to detect stat output format. All other listener/server/backend corresponding functions already have it.
2024-04-16 16:17:48 +00:00
if (!stats_fill_fe_line(px, 0, stats, ST_I_PX_MAX, &(ctx->field_num)))
MINOR: contrib/prometheus-exporter: use fill_fe_stats for frontend dump use `stats_fill_fe_stats` when possible to avoid duplicating code; make use of field selector to get the needed field only. this should not introduce any difference of output. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-17 17:27:46 +00:00
return -1;
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
switch (ctx->field_num) {
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_STATUS:
MINOR: proxy: Introduce proxy flags to replace disabled bitfield This change is required to support TCP/HTTP rules in defaults sections. The 'disabled' bitfield in the proxy structure, used to know if a proxy is disabled or stopped, is replaced a generic bitfield named 'flags'. PR_DISABLED and PR_STOPPED flags are renamed to PR_FL_DISABLED and PR_FL_STOPPED respectively. In addition, everywhere there is a test to know if a proxy is disabled or stopped, there is now a bitwise AND operation on PR_FL_DISABLED and/or PR_FL_STOPPED flags.
2021-10-06 12:24:19 +00:00
state = !(px->flags & PR_FL_STOPPED);
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
for (; ctx->obj_state < PROMEX_FRONT_STATE_COUNT; ctx->obj_state++) {
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[1].name = ist("state");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
labels[1].value = promex_front_st[ctx->obj_state];
val = mkf_u32(FO_STATUS, state == ctx->obj_state);
MINOR: promex: Add a param to override the description when a metric is dumped When a metric is dumped, it is now possible to specify a custom description. We will add the support for extra counters. The list of these counters is retrived dynamically. Thus the description must be dynamic too. Note it was already possible to customize the metric name.
2024-01-29 15:37:35 +00:00
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for frontend/backend/servers states to labels values. the main motivation being I realised it is very difficult to make use of it without hard coded quirks on prometheus client side; especially because the main use is often to group by state, which is harder when the state is the value of the metric. in order to achieve that we iterate on the status metric to generate labels, and so as many metrics. this is the first step to resolve github issue #1029 A second step should address health check states. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:17 +00:00
goto full;
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->obj_state = 0;
MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for frontend/backend/servers states to labels values. the main motivation being I realised it is very difficult to make use of it without hard coded quirks on prometheus client side; especially because the main use is often to group by state, which is harder when the state is the value of the metric. in order to achieve that we iterate on the status metric to generate labels, and so as many metrics. this is the first step to resolve github issue #1029 A second step should address health check states. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:17 +00:00
goto next_px;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_REQ_RATE_MAX:
case ST_I_PX_REQ_TOT:
case ST_I_PX_INTERCEPTED:
case ST_I_PX_CACHE_LOOKUPS:
case ST_I_PX_CACHE_HITS:
case ST_I_PX_COMP_IN:
case ST_I_PX_COMP_OUT:
case ST_I_PX_COMP_BYP:
case ST_I_PX_COMP_RSP:
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
if (px->mode != PR_MODE_HTTP)
goto next_px;
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_HRSP_1XX:
case ST_I_PX_HRSP_2XX:
case ST_I_PX_HRSP_3XX:
case ST_I_PX_HRSP_4XX:
case ST_I_PX_HRSP_5XX:
case ST_I_PX_HRSP_OTHER:
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
if (px->mode != PR_MODE_HTTP)
goto next_px;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
if (ctx->field_num != ST_I_PX_HRSP_1XX)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~PROMEX_FL_METRIC_HDR;
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[1].name = ist("code");
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
labels[1].value = promex_hrsp_code[ctx->field_num - ST_I_PX_HRSP_1XX];
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
default:
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto full;
next_px:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = px->next;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_METRIC_HDR;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
/* Skip extra counters */
if (!(ctx->flags & PROMEX_FL_EXTRA_COUNTERS))
goto end;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!mod) {
mod = LIST_NEXT(&stats_module_list[STATS_DOMAIN_PROXY], typeof(mod), list);
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
ctx->mod_field_num = 0;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
list_for_each_entry_from(mod, &stats_module_list[STATS_DOMAIN_PROXY], list) {
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
void *counters;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!(stats_px_get_cap(mod->domain_flags) & STATS_PX_CAP_FE)) {
ctx->field_num += mod->stats_count;
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
ctx->mod_field_num = 0;
continue;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
for (;ctx->mod_field_num < mod->stats_count; ctx->mod_field_num++) {
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
name = ist2(mod->stats[ctx->mod_field_num].name, strlen(mod->stats[ctx->mod_field_num].name));
desc = ist2(mod->stats[ctx->mod_field_num].desc, strlen(mod->stats[ctx->mod_field_num].desc));
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, name))
continue;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!px)
px = proxies_list;
while (px) {
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
struct promex_label labels[PROMEX_MAX_LABELS-1] = {};
struct promex_metric metric;
labels[0].name = ist("proxy");
labels[0].value = ist2(px->id, strlen(px->id));
labels[1].name = ist("mod");
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
labels[1].value = ist2(mod->name, strlen(mod->name));
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
/* skip the disabled proxies, global frontend and non-networked ones */
if ((px->flags & PR_FL_DISABLED) || px->uuid <= 0 || !(px->cap & PR_CAP_FE))
goto next_px2;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
counters = EXTRA_COUNTERS_GET(px->extra_counters_fe, mod);
if (!mod->fill_stats(counters, stats + ctx->field_num, &ctx->mod_field_num))
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
return -1;
val = stats[ctx->field_num + ctx->mod_field_num];
metric.type = ((val.type == FN_GAUGE) ? PROMEX_MT_GAUGE : PROMEX_MT_COUNTER);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc, metric.type,
&val, labels, &out, max))
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
goto full;
next_px2:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = px->next;
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
}
ctx->flags |= PROMEX_FL_METRIC_HDR;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
ctx->field_num += mod->stats_count;
MEDIUM: promex: Dump frontends extra counters if requested Dump extra counter for the frontends. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:24 +00:00
ctx->mod_field_num = 0;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = NULL;
mod = NULL;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
end:
BUG/MINOR: contrib/prometheus-exporter: Don't try to add empty data blocks When the response buffer is full and nothing more can be inserted, it is important to not try to insert an empty data block. Otherwise, when the function channel_add_input() is called, the flag CF_READ_PARTIAL is set on the response channel while nothing was read and the stream is uselessly woken up. Finally, we have loop while the response buffer is full. This patch must be backported to 2.0.
2019-07-04 08:03:28 +00:00
if (out.len) {
if (!htx_add_data_atonce(htx, out))
return -1; /* Unexpected and unrecoverable error */
channel_add_input(chn, out.len);
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
/* Save pointers (0=current proxy, 1=current stats module) of the current context */
ctx->p[0] = px;
ctx->p[1] = mod;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return ret;
full:
ret = 0;
goto end;
}
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
/* Dump listener metrics (prefixed by "haproxy_listen_"). It returns 1 on
* success, 0 if <htx> is full and -1 in case of any error. */
static int promex_dump_listener_metrics(struct appctx *appctx, struct htx *htx)
{
static struct ist prefix = IST("haproxy_listener_");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
struct promex_ctx *ctx = appctx->svcctx;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
struct proxy *px = ctx->p[0];
struct listener *li = ctx->p[1];
struct stats_module *mod = ctx->p[2];
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
struct field val;
CLEANUP: applet: rename appctx_cs() to appctx_sc() It returns a stream connector, not a conn_stream anymore, so let's fix its name.
2022-05-27 09:08:15 +00:00
struct channel *chn = sc_ic(appctx_sc(appctx));
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
struct ist name, desc, out = ist2(trash.area, 0);
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
size_t max = htx_get_max_blksz(htx, channel_htx_recv_max(chn, htx));
MINOR: stats: rename ambiguous stat_l and stat_count Statistics were extended with the introduction of stats module. This mechanism allows to expose various metrics for several haproxy components. As a consequence of this, some static variables were transformed to dynamic ones to be able to regroup all statistics definition. Rename these variables with more explicit naming : * stat_lines can be used to generate one line of statistics for any module using struct field as value * metrics and metrics_len are used to stored description of metrics indexed by module Note that info is not integrated in the statistics module mechanism. However, it could be done in the future to better reflect its purpose.
2024-04-22 09:19:17 +00:00
struct field *stats = stat_lines[STATS_DOMAIN_PROXY];
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
int ret = 1;
enum li_status status;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
for (;ctx->field_num < ST_I_PX_MAX; ctx->field_num++) {
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
if (!(promex_st_metrics[ctx->field_num].flags & ctx->flags))
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
continue;
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
name = promex_st_li_metrics_names[ctx->field_num];
desc = promex_st_metric_desc[ctx->field_num];
if (!isttest(name))
name = promex_st_metrics[ctx->field_num].n;
if (!isttest(desc))
MINOR: stats: update ambiguous "metrics" naming to "stat_cols" The name "metrics" was chosen to represent the various list of haproxy exposed statistics. However, it is deemed as ambiguous as some stats are indeed metric in the true sense, but some are not, as highlighted by various "enum field_origin" values. Replace it by the new name "stat_cols" for statistic columns. Along with the already existing notion of stat lines it should better reflect its purpose.
2024-04-23 09:06:00 +00:00
desc = ist(stat_cols_px[ctx->field_num].desc);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, name))
continue;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!px)
px = proxies_list;
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
while (px) {
struct promex_label labels[PROMEX_MAX_LABELS-1] = {};
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
labels[0].name = ist("proxy");
labels[0].value = ist2(px->id, strlen(px->id));
/* skip the disabled proxies, global frontend and non-networked ones */
MINOR: proxy: Introduce proxy flags to replace disabled bitfield This change is required to support TCP/HTTP rules in defaults sections. The 'disabled' bitfield in the proxy structure, used to know if a proxy is disabled or stopped, is replaced a generic bitfield named 'flags'. PR_DISABLED and PR_STOPPED flags are renamed to PR_FL_DISABLED and PR_FL_STOPPED respectively. In addition, everywhere there is a test to know if a proxy is disabled or stopped, there is now a bitwise AND operation on PR_FL_DISABLED and/or PR_FL_STOPPED flags.
2021-10-06 12:24:19 +00:00
if ((px->flags & PR_FL_DISABLED) || px->uuid <= 0 || !(px->cap & PR_CAP_FE))
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
goto next_px;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!li)
li = LIST_NEXT(&px->conf.listeners, struct listener *, by_fe);
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
list_for_each_entry_from(li, &px->conf.listeners, by_fe) {
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
if (!li->counters)
continue;
BUG/MEDIUM: contrib/prometheus-exporter: fix segfault in listener name dump We need to check whether listener is empty before doing anything; in that case, we were trying to dump listerner name while name is null. So simply move the counters check above, which validate all possible cases when the listener is empty. This is very similar to what is done in stats.c see also the trace: Thread 1 "haproxy" received signal SIGSEGV, Segmentation fault. __strlen_sse2 () at ../sysdeps/x86_64/multiarch/../strlen.S:120 120     ../sysdeps/x86_64/multiarch/../strlen.S: No such file or directory. (gdb) bt #0  __strlen_sse2 () at ../sysdeps/x86_64/multiarch/../strlen.S:120 #1  0x00005555555b716b in promex_dump_listener_metrics (htx=0x5555558fadf0, appctx=0x555555926070) at contrib/prometheus-exporter/service-prometheus.c:722 #2  promex_dump_metrics (htx=0x5555558fadf0, si=0x555555925920, appctx=0x555555926070) at contrib/prometheus-exporter/service-prometheus.c:1200 #3  promex_appctx_handle_io (appctx=0x555555926070) at contrib/prometheus-exporter/service-prometheus.c:1477 #4  0x00005555556f0c94 in task_run_applet (t=0x555555926180, context=0x555555926070, state=<optimized out>) at src/applet.c:88 #5  0x00005555556bc6d8 in run_tasks_from_lists (budgets=budgets@entry=0x7fffffffe374) at src/task.c:548 #6  0x00005555556bd1a0 in process_runnable_tasks () at src/task.c:750 #7  0x0000555555696cdd in run_poll_loop () at src/haproxy.c:2870 #8  0x0000555555697025 in run_thread_poll_loop (data=data@entry=0x0) at src/haproxy.c:3035 #9  0x0000555555596c90 in main (argc=<optimized out>, argv=0x7fffffffe818) at src/haproxy.c:3723 quit) this bug was introduced by commit e3f7bd5ae9e969cbfe87e4130d06bff7a3e814c6 ("MEDIUM: contrib/prometheus-exporter: add listen stats"), which is present for 2.4 only, so no backport needed. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-24 23:53:13 +00:00
labels[1].name = ist("listener");
labels[1].value = ist2(li->name, strlen(li->name));
MINOR: stats: use stricter naming stats/field/line Several unique names were used for different purposes under statistics implementation. This caused the code to be difficult to understand. * stat/stats name is removed when a more specific name could be used * restrict field usage to purely refer to <struct field> which represents a raw stat value. * use "line" naming to represent an array of <struct field>
2024-04-22 08:26:23 +00:00
if (!stats_fill_li_line(px, li, 0, stats,
ST_I_PX_MAX, &(ctx->field_num)))
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
return -1;
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
switch (ctx->field_num) {
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_STATUS:
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
status = get_li_status(li);
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
for (; ctx->obj_state < LI_STATE_COUNT; ctx->obj_state++) {
val = mkf_u32(FO_STATUS, status == ctx->obj_state);
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
labels[2].name = ist("state");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
labels[2].value = ist(li_status_st[ctx->obj_state]);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
goto full;
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->obj_state = 0;
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
continue;
default:
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
}
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
goto full;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
li = NULL;
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
next_px:
px = px->next;
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_METRIC_HDR;
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
}
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
/* Skip extra counters */
if (!(ctx->flags & PROMEX_FL_EXTRA_COUNTERS))
goto end;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!mod) {
mod = LIST_NEXT(&stats_module_list[STATS_DOMAIN_PROXY], typeof(mod), list);
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
ctx->mod_field_num = 0;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
list_for_each_entry_from(mod, &stats_module_list[STATS_DOMAIN_PROXY], list) {
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
void *counters;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!(stats_px_get_cap(mod->domain_flags) & STATS_PX_CAP_LI)) {
ctx->field_num += mod->stats_count;
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
ctx->mod_field_num = 0;
continue;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
for (;ctx->mod_field_num < mod->stats_count; ctx->mod_field_num++) {
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
name = ist2(mod->stats[ctx->mod_field_num].name, strlen(mod->stats[ctx->mod_field_num].name));
desc = ist2(mod->stats[ctx->mod_field_num].desc, strlen(mod->stats[ctx->mod_field_num].desc));
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, name))
continue;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!px)
px = proxies_list;
while (px) {
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
struct promex_label labels[PROMEX_MAX_LABELS-1] = {};
struct promex_metric metric;
labels[0].name = ist("proxy");
labels[0].value = ist2(px->id, strlen(px->id));
/* skip the disabled proxies, global frontend and non-networked ones */
if ((px->flags & PR_FL_DISABLED) || px->uuid <= 0 || !(px->cap & PR_CAP_FE))
goto next_px2;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!li)
li = LIST_NEXT(&px->conf.listeners, struct listener *, by_fe);
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
list_for_each_entry_from(li, &px->conf.listeners, by_fe) {
if (!li->counters)
continue;
labels[1].name = ist("listener");
labels[1].value = ist2(li->name, strlen(li->name));
labels[2].name = ist("mod");
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
labels[2].value = ist2(mod->name, strlen(mod->name));
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
counters = EXTRA_COUNTERS_GET(li->extra_counters, mod);
if (!mod->fill_stats(counters, stats + ctx->field_num, &ctx->mod_field_num))
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
return -1;
val = stats[ctx->field_num + ctx->mod_field_num];
metric.type = ((val.type == FN_GAUGE) ? PROMEX_MT_GAUGE : PROMEX_MT_COUNTER);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc, metric.type,
&val, labels, &out, max))
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
goto full;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
li = NULL;
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
next_px2:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = px->next;
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
}
ctx->flags |= PROMEX_FL_METRIC_HDR;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
ctx->field_num += mod->stats_count;
MEDIUM: promex: Dump listeners extra counters if requested Dump extra counter for the listeners. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:42:02 +00:00
ctx->mod_field_num = 0;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = NULL;
li = NULL;
mod = NULL;
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
end:
if (out.len) {
if (!htx_add_data_atonce(htx, out))
return -1; /* Unexpected and unrecoverable error */
channel_add_input(chn, out.len);
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
/* Save pointers (0=current proxy, 1=current listener, 2=current stats module) of the current context */
ctx->p[0] = px;
ctx->p[1] = li;
ctx->p[2] = mod;
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
return ret;
full:
ret = 0;
goto end;
}
CLEANUP: assorted typo fixes in the code and comments This is fifth iteration of typo fixes
2020-03-23 17:28:40 +00:00
/* Dump backends metrics (prefixed by "haproxy_backend_"). It returns 1 on success,
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
* 0 if <htx> is full and -1 in case of any error. */
static int promex_dump_back_metrics(struct appctx *appctx, struct htx *htx)
{
static struct ist prefix = IST("haproxy_backend_");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
struct promex_ctx *ctx = appctx->svcctx;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
struct proxy *px = ctx->p[0];
struct stats_module *mod = ctx->p[2];
MINOR: promex: backend aggregated server check status - add new metric: `haproxy_backend_agg_server_check_status` it counts the number of servers matching a specific check status this permits to exclude per server check status as the usage is often to rely on the total. Indeed in large setup having thousands of servers per backend the memory impact is not neglible to store the per server metric. - realign promex_str_metrics array quite simple implementation - we could improve it later by adding an internal state to the prometheus exporter, thus to avoid counting at every dump. this patch is an attempt to close github issue #1312. It may bebackported to 2.4 if requested. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-11-07 09:18:47 +00:00
struct server *sv;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
struct field val;
CLEANUP: applet: rename appctx_cs() to appctx_sc() It returns a stream connector, not a conn_stream anymore, so let's fix its name.
2022-05-27 09:08:15 +00:00
struct channel *chn = sc_ic(appctx_sc(appctx));
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
struct ist name, desc, out = ist2(trash.area, 0);
BUG/MINOR: contrib/prometheus-exporter: Respect the reserve when data are sent The previous commit e6cdfe574 ("BUG/MINOR: contrib/prometheus-exporter: Don't use channel_htx_recv_max()") is buggy. The buffer's reserve must be respected. This patch must be backported to 2.0 and 1.9.
2019-07-03 09:43:17 +00:00
size_t max = htx_get_max_blksz(htx, channel_htx_recv_max(chn, htx));
MINOR: stats: rename ambiguous stat_l and stat_count Statistics were extended with the introduction of stats module. This mechanism allows to expose various metrics for several haproxy components. As a consequence of this, some static variables were transformed to dynamic ones to be able to regroup all statistics definition. Rename these variables with more explicit naming : * stat_lines can be used to generate one line of statistics for any module using struct field as value * metrics and metrics_len are used to stored description of metrics indexed by module Note that info is not integrated in the statistics module mechanism. However, it could be done in the future to better reflect its purpose.
2024-04-22 09:19:17 +00:00
struct field *stats = stat_lines[STATS_DOMAIN_PROXY];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
int ret = 1;
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
double secs;
MINOR: promex: backend aggregated server check status - add new metric: `haproxy_backend_agg_server_check_status` it counts the number of servers matching a specific check status this permits to exclude per server check status as the usage is often to rely on the total. Indeed in large setup having thousands of servers per backend the memory impact is not neglible to store the per server metric. - realign promex_str_metrics array quite simple implementation - we could improve it later by adding an internal state to the prometheus exporter, thus to avoid counting at every dump. this patch is an attempt to close github issue #1312. It may bebackported to 2.4 if requested. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-11-07 09:18:47 +00:00
enum promex_back_state bkd_state;
enum promex_srv_state srv_state;
MINOR: promex: introduce haproxy_backend_agg_check_status This patch introduces haproxy_backend_agg_check_status metric as we wanted in 42d7c402d but with the right data source. This patch could be backported as far as 2.4.
2022-12-08 09:17:01 +00:00
enum healthcheck_status srv_check_status;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
for (;ctx->field_num < ST_I_PX_MAX; ctx->field_num++) {
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
if (!(promex_st_metrics[ctx->field_num].flags & ctx->flags))
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
continue;
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
name = promex_st_back_metrics_names[ctx->field_num];
desc = promex_st_metric_desc[ctx->field_num];
if (!isttest(name))
name = promex_st_metrics[ctx->field_num].n;
if (!isttest(desc))
MINOR: stats: update ambiguous "metrics" naming to "stat_cols" The name "metrics" was chosen to represent the various list of haproxy exposed statistics. However, it is deemed as ambiguous as some stats are indeed metric in the true sense, but some are not, as highlighted by various "enum field_origin" values. Replace it by the new name "stat_cols" for statistic columns. Along with the already existing notion of stat lines it should better reflect its purpose.
2024-04-23 09:06:00 +00:00
desc = ist(stat_cols_px[ctx->field_num].desc);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, name))
continue;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!px)
px = proxies_list;
while (px) {
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
struct promex_label labels[PROMEX_MAX_LABELS-1] = {};
MINOR: promex: backend aggregated server check status - add new metric: `haproxy_backend_agg_server_check_status` it counts the number of servers matching a specific check status this permits to exclude per server check status as the usage is often to rely on the total. Indeed in large setup having thousands of servers per backend the memory impact is not neglible to store the per server metric. - realign promex_str_metrics array quite simple implementation - we could improve it later by adding an internal state to the prometheus exporter, thus to avoid counting at every dump. this patch is an attempt to close github issue #1312. It may bebackported to 2.4 if requested. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-11-07 09:18:47 +00:00
unsigned int srv_state_count[PROMEX_SRV_STATE_COUNT] = { 0 };
MINOR: promex: introduce haproxy_backend_agg_check_status This patch introduces haproxy_backend_agg_check_status metric as we wanted in 42d7c402d but with the right data source. This patch could be backported as far as 2.4.
2022-12-08 09:17:01 +00:00
unsigned int srv_check_count[HCHK_STATUS_SIZE] = { 0 };
const char *check_state;
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[0].name = ist("proxy");
labels[0].value = ist2(px->id, strlen(px->id));
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
/* skip the disabled proxies, global frontend and non-networked ones */
MINOR: proxy: Introduce proxy flags to replace disabled bitfield This change is required to support TCP/HTTP rules in defaults sections. The 'disabled' bitfield in the proxy structure, used to know if a proxy is disabled or stopped, is replaced a generic bitfield named 'flags'. PR_DISABLED and PR_STOPPED flags are renamed to PR_FL_DISABLED and PR_FL_STOPPED respectively. In addition, everywhere there is a test to know if a proxy is disabled or stopped, there is now a bitwise AND operation on PR_FL_DISABLED and/or PR_FL_STOPPED flags.
2021-10-06 12:24:19 +00:00
if ((px->flags & PR_FL_DISABLED) || px->uuid <= 0 || !(px->cap & PR_CAP_BE))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto next_px;
MINOR: stats: use stricter naming stats/field/line Several unique names were used for different purposes under statistics implementation. This caused the code to be difficult to understand. * stat/stats name is removed when a more specific name could be used * restrict field usage to purely refer to <struct field> which represents a raw stat value. * use "line" naming to represent an array of <struct field>
2024-04-22 08:26:23 +00:00
if (!stats_fill_be_line(px, 0, stats, ST_I_PX_MAX, &(ctx->field_num)))
MINOR: contrib/prometheus-exporter: use fill_be_stats for backend dump use `stats_fill_be_stats` when possible to avoid duplicating code; make use of field selector to get the needed field only. the only difference is on `haproxy_backend_downtime_seconds_total` as stats.c is testing `px->srv`. This behaviour is present since commit 7344f4789321ef8ce2ce17cf73dabd672f7c8c6 ("MINOR: stats: only report backend's down time if it has servers"). The end result is a NaN instead of a zero when no server are present. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-25 16:29:02 +00:00
return -1;
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
switch (ctx->field_num) {
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_AGG_SRV_CHECK_STATUS: // DEPRECATED
case ST_I_PX_AGG_SRV_STATUS:
MINOR: promex: backend aggregated server check status - add new metric: `haproxy_backend_agg_server_check_status` it counts the number of servers matching a specific check status this permits to exclude per server check status as the usage is often to rely on the total. Indeed in large setup having thousands of servers per backend the memory impact is not neglible to store the per server metric. - realign promex_str_metrics array quite simple implementation - we could improve it later by adding an internal state to the prometheus exporter, thus to avoid counting at every dump. this patch is an attempt to close github issue #1312. It may bebackported to 2.4 if requested. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-11-07 09:18:47 +00:00
if (!px->srv)
goto next_px;
sv = px->srv;
while (sv) {
srv_state = promex_srv_status(sv);
srv_state_count[srv_state] += 1;
sv = sv->next;
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
for (; ctx->obj_state < PROMEX_SRV_STATE_COUNT; ctx->obj_state++) {
val = mkf_u32(FN_GAUGE, srv_state_count[ctx->obj_state]);
MINOR: promex: backend aggregated server check status - add new metric: `haproxy_backend_agg_server_check_status` it counts the number of servers matching a specific check status this permits to exclude per server check status as the usage is often to rely on the total. Indeed in large setup having thousands of servers per backend the memory impact is not neglible to store the per server metric. - realign promex_str_metrics array quite simple implementation - we could improve it later by adding an internal state to the prometheus exporter, thus to avoid counting at every dump. this patch is an attempt to close github issue #1312. It may bebackported to 2.4 if requested. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-11-07 09:18:47 +00:00
labels[1].name = ist("state");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
labels[1].value = promex_srv_st[ctx->obj_state];
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
MINOR: promex: introduce haproxy_backend_agg_check_status This patch introduces haproxy_backend_agg_check_status metric as we wanted in 42d7c402d but with the right data source. This patch could be backported as far as 2.4.
2022-12-08 09:17:01 +00:00
goto full;
}
ctx->obj_state = 0;
goto next_px;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_AGG_CHECK_STATUS:
MINOR: promex: introduce haproxy_backend_agg_check_status This patch introduces haproxy_backend_agg_check_status metric as we wanted in 42d7c402d but with the right data source. This patch could be backported as far as 2.4.
2022-12-08 09:17:01 +00:00
if (!px->srv)
goto next_px;
sv = px->srv;
while (sv) {
BUG/MINOR: promex: fix backend_agg_check_status When a server is in maintenance, the check.status is no longer updated. Therefore, we shouldn't consider check.status if the checks are not active. This check was properly implemented in the haproxy_server_check_status metric, but wasn't carried over to backend_agg_check_status, which introduced inconsistencies between the two metrics. [cf: This patch must be backported as far as 2.4]
2023-09-12 09:37:55 +00:00
if ((sv->check.state & (CHK_ST_ENABLED|CHK_ST_PAUSED)) == CHK_ST_ENABLED) {
srv_check_status = sv->check.status;
srv_check_count[srv_check_status] += 1;
}
MINOR: promex: introduce haproxy_backend_agg_check_status This patch introduces haproxy_backend_agg_check_status metric as we wanted in 42d7c402d but with the right data source. This patch could be backported as far as 2.4.
2022-12-08 09:17:01 +00:00
sv = sv->next;
}
for (; ctx->obj_state < HCHK_STATUS_SIZE; ctx->obj_state++) {
if (get_check_status_result(ctx->obj_state) < CHK_RES_FAILED)
continue;
val = mkf_u32(FO_STATUS, srv_check_count[ctx->obj_state]);
check_state = get_check_status_info(ctx->obj_state);
labels[1].name = ist("state");
labels[1].value = ist(check_state);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
MINOR: promex: backend aggregated server check status - add new metric: `haproxy_backend_agg_server_check_status` it counts the number of servers matching a specific check status this permits to exclude per server check status as the usage is often to rely on the total. Indeed in large setup having thousands of servers per backend the memory impact is not neglible to store the per server metric. - realign promex_str_metrics array quite simple implementation - we could improve it later by adding an internal state to the prometheus exporter, thus to avoid counting at every dump. this patch is an attempt to close github issue #1312. It may bebackported to 2.4 if requested. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-11-07 09:18:47 +00:00
goto full;
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->obj_state = 0;
MINOR: promex: backend aggregated server check status - add new metric: `haproxy_backend_agg_server_check_status` it counts the number of servers matching a specific check status this permits to exclude per server check status as the usage is often to rely on the total. Indeed in large setup having thousands of servers per backend the memory impact is not neglible to store the per server metric. - realign promex_str_metrics array quite simple implementation - we could improve it later by adding an internal state to the prometheus exporter, thus to avoid counting at every dump. this patch is an attempt to close github issue #1312. It may bebackported to 2.4 if requested. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-11-07 09:18:47 +00:00
goto next_px;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_STATUS:
MINOR: promex: backend aggregated server check status - add new metric: `haproxy_backend_agg_server_check_status` it counts the number of servers matching a specific check status this permits to exclude per server check status as the usage is often to rely on the total. Indeed in large setup having thousands of servers per backend the memory impact is not neglible to store the per server metric. - realign promex_str_metrics array quite simple implementation - we could improve it later by adding an internal state to the prometheus exporter, thus to avoid counting at every dump. this patch is an attempt to close github issue #1312. It may bebackported to 2.4 if requested. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-11-07 09:18:47 +00:00
bkd_state = ((px->lbprm.tot_weight > 0 || !px->srv) ? 1 : 0);
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
for (; ctx->obj_state < PROMEX_BACK_STATE_COUNT; ctx->obj_state++) {
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[1].name = ist("state");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
labels[1].value = promex_back_st[ctx->obj_state];
val = mkf_u32(FO_STATUS, bkd_state == ctx->obj_state);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for frontend/backend/servers states to labels values. the main motivation being I realised it is very difficult to make use of it without hard coded quirks on prometheus client side; especially because the main use is often to group by state, which is harder when the state is the value of the metric. in order to achieve that we iterate on the status metric to generate labels, and so as many metrics. this is the first step to resolve github issue #1029 A second step should address health check states. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:17 +00:00
goto full;
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->obj_state = 0;
MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for frontend/backend/servers states to labels values. the main motivation being I realised it is very difficult to make use of it without hard coded quirks on prometheus client side; especially because the main use is often to group by state, which is harder when the state is the value of the metric. in order to achieve that we iterate on the status metric to generate labels, and so as many metrics. this is the first step to resolve github issue #1029 A second step should address health check states. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:17 +00:00
goto next_px;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_QTIME:
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
secs = (double)swrate_avg(px->be_counters.q_time, TIME_STATS_SAMPLES) / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_AVG, secs);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_CTIME:
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
secs = (double)swrate_avg(px->be_counters.c_time, TIME_STATS_SAMPLES) / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_AVG, secs);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_RTIME:
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
secs = (double)swrate_avg(px->be_counters.d_time, TIME_STATS_SAMPLES) / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_AVG, secs);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_TTIME:
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
secs = (double)swrate_avg(px->be_counters.t_time, TIME_STATS_SAMPLES) / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_AVG, secs);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_QT_MAX:
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
secs = (double)px->be_counters.qtime_max / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_MAX, secs);
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_CT_MAX:
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
secs = (double)px->be_counters.ctime_max / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_MAX, secs);
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_RT_MAX:
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
secs = (double)px->be_counters.dtime_max / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_MAX, secs);
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_TT_MAX:
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
secs = (double)px->be_counters.ttime_max / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_MAX, secs);
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_REQ_TOT:
case ST_I_PX_CACHE_LOOKUPS:
case ST_I_PX_CACHE_HITS:
case ST_I_PX_COMP_IN:
case ST_I_PX_COMP_OUT:
case ST_I_PX_COMP_BYP:
case ST_I_PX_COMP_RSP:
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
if (px->mode != PR_MODE_HTTP)
goto next_px;
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_HRSP_1XX:
case ST_I_PX_HRSP_2XX:
case ST_I_PX_HRSP_3XX:
case ST_I_PX_HRSP_4XX:
case ST_I_PX_HRSP_5XX:
case ST_I_PX_HRSP_OTHER:
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
if (px->mode != PR_MODE_HTTP)
goto next_px;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
if (ctx->field_num != ST_I_PX_HRSP_1XX)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~PROMEX_FL_METRIC_HDR;
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[1].name = ist("code");
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
labels[1].value = promex_hrsp_code[ctx->field_num - ST_I_PX_HRSP_1XX];
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
default:
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto full;
next_px:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = px->next;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_METRIC_HDR;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
/* Skip extra counters */
if (!(ctx->flags & PROMEX_FL_EXTRA_COUNTERS))
goto end;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!mod) {
mod = LIST_NEXT(&stats_module_list[STATS_DOMAIN_PROXY], typeof(mod), list);
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
ctx->mod_field_num = 0;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
list_for_each_entry_from(mod, &stats_module_list[STATS_DOMAIN_PROXY], list) {
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
void *counters;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!(stats_px_get_cap(mod->domain_flags) & STATS_PX_CAP_BE)) {
ctx->field_num += mod->stats_count;
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
ctx->mod_field_num = 0;
continue;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
for (;ctx->mod_field_num < mod->stats_count; ctx->mod_field_num++) {
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
name = ist2(mod->stats[ctx->mod_field_num].name, strlen(mod->stats[ctx->mod_field_num].name));
desc = ist2(mod->stats[ctx->mod_field_num].desc, strlen(mod->stats[ctx->mod_field_num].desc));
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, name))
continue;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!px)
px = proxies_list;
while (px) {
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
struct promex_label labels[PROMEX_MAX_LABELS-1] = {};
struct promex_metric metric;
labels[0].name = ist("proxy");
labels[0].value = ist2(px->id, strlen(px->id));
labels[1].name = ist("mod");
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
labels[1].value = ist2(mod->name, strlen(mod->name));
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
/* skip the disabled proxies, global frontend and non-networked ones */
if ((px->flags & PR_FL_DISABLED) || px->uuid <= 0 || !(px->cap & PR_CAP_BE))
goto next_px2;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
counters = EXTRA_COUNTERS_GET(px->extra_counters_be, mod);
if (!mod->fill_stats(counters, stats + ctx->field_num, &ctx->mod_field_num))
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
return -1;
val = stats[ctx->field_num + ctx->mod_field_num];
metric.type = ((val.type == FN_GAUGE) ? PROMEX_MT_GAUGE : PROMEX_MT_COUNTER);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc, metric.type,
&val, labels, &out, max))
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
goto full;
next_px2:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = px->next;
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
}
ctx->flags |= PROMEX_FL_METRIC_HDR;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
ctx->field_num += mod->stats_count;
MEDIUM: promex: Dump backends extra counters if requested Dump extra counter for the backend. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:42 +00:00
ctx->mod_field_num = 0;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = NULL;
mod = NULL;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
end:
BUG/MINOR: contrib/prometheus-exporter: Don't try to add empty data blocks When the response buffer is full and nothing more can be inserted, it is important to not try to insert an empty data block. Otherwise, when the function channel_add_input() is called, the flag CF_READ_PARTIAL is set on the response channel while nothing was read and the stream is uselessly woken up. Finally, we have loop while the response buffer is full. This patch must be backported to 2.0.
2019-07-04 08:03:28 +00:00
if (out.len) {
if (!htx_add_data_atonce(htx, out))
return -1; /* Unexpected and unrecoverable error */
channel_add_input(chn, out.len);
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
/* Save pointers (0=current proxy, 1=current stats module) of the current context */
ctx->p[0] = px;
ctx->p[1] = mod;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return ret;
full:
ret = 0;
goto end;
}
CLEANUP: assorted typo fixes in the code and comments This is fifth iteration of typo fixes
2020-03-23 17:28:40 +00:00
/* Dump servers metrics (prefixed by "haproxy_server_"). It returns 1 on success,
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
* 0 if <htx> is full and -1 in case of any error. */
static int promex_dump_srv_metrics(struct appctx *appctx, struct htx *htx)
{
static struct ist prefix = IST("haproxy_server_");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
struct promex_ctx *ctx = appctx->svcctx;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
struct proxy *px = ctx->p[0];
struct server *sv = ctx->p[1];
struct stats_module *mod = ctx->p[2];
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
struct field val;
CLEANUP: applet: rename appctx_cs() to appctx_sc() It returns a stream connector, not a conn_stream anymore, so let's fix its name.
2022-05-27 09:08:15 +00:00
struct channel *chn = sc_ic(appctx_sc(appctx));
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
struct ist name, desc, out = ist2(trash.area, 0);
BUG/MINOR: contrib/prometheus-exporter: Respect the reserve when data are sent The previous commit e6cdfe574 ("BUG/MINOR: contrib/prometheus-exporter: Don't use channel_htx_recv_max()") is buggy. The buffer's reserve must be respected. This patch must be backported to 2.0 and 1.9.
2019-07-03 09:43:17 +00:00
size_t max = htx_get_max_blksz(htx, channel_htx_recv_max(chn, htx));
MINOR: stats: rename ambiguous stat_l and stat_count Statistics were extended with the introduction of stats module. This mechanism allows to expose various metrics for several haproxy components. As a consequence of this, some static variables were transformed to dynamic ones to be able to regroup all statistics definition. Rename these variables with more explicit naming : * stat_lines can be used to generate one line of statistics for any module using struct field as value * metrics and metrics_len are used to stored description of metrics indexed by module Note that info is not integrated in the statistics module mechanism. However, it could be done in the future to better reflect its purpose.
2024-04-22 09:19:17 +00:00
struct field *stats = stat_lines[STATS_DOMAIN_PROXY];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
int ret = 1;
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
double secs;
MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for frontend/backend/servers states to labels values. the main motivation being I realised it is very difficult to make use of it without hard coded quirks on prometheus client side; especially because the main use is often to group by state, which is harder when the state is the value of the metric. in order to achieve that we iterate on the status metric to generate labels, and so as many metrics. this is the first step to resolve github issue #1029 A second step should address health check states. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:17 +00:00
enum promex_srv_state state;
MAJOR: contrib/prometheus-exporter: move health check status to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for health check states to labels values. The diff is quite small thanks to the preparation work from Christopher to allow more flexibility in labels, see commit 5a2f938732126f43bbf4cea5482c01552b0d0314 ("MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones") this is a follow up of commit c6464591a365bfcf509b322bdaa4d608c9395d75 ("MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels"). The main goal being to be better aligned with prometheus use cases in terms of queries. More specifically to health checks, Pierre C. mentioned the possible quirks he had to put in place in order to make use of those metrics through prometheus: <aggregator_function> by(proxy, check_status) (count_values by(proxy, instance) ("check_status", haproxy_server_check_status)) I am perfectly aware this introduces a lot more metrics but I don't see how we can improve the usability without it. The main issue remains in the cardinality of the states which are > 20. Prometheus recommends to stay below a cardinality of 10 for a given metric but I consider our case very specific, because highly linked to the level of precision haproxy exposes. Even before this patch I saw several large production setup (a few hundreds of MB in output) which are making use of the scope parameter to simply ignore the server metrics, so that the scrapping can be faster, and memory consumed on client side not too high. So I believe we should eventually continue in that direction and offer more granularity of filtering of the output. That being said it is already possible to filter out the data on prometheus client side. this is related to github issue #1029 Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-01 12:11:51 +00:00
const char *check_state;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
for (;ctx->field_num < ST_I_PX_MAX; ctx->field_num++) {
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
if (!(promex_st_metrics[ctx->field_num].flags & ctx->flags))
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
continue;
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
name = promex_st_srv_metrics_names[ctx->field_num];
desc = promex_st_metric_desc[ctx->field_num];
if (!isttest(name))
name = promex_st_metrics[ctx->field_num].n;
if (!isttest(desc))
MINOR: stats: update ambiguous "metrics" naming to "stat_cols" The name "metrics" was chosen to represent the various list of haproxy exposed statistics. However, it is deemed as ambiguous as some stats are indeed metric in the true sense, but some are not, as highlighted by various "enum field_origin" values. Replace it by the new name "stat_cols" for statistic columns. Along with the already existing notion of stat lines it should better reflect its purpose.
2024-04-23 09:06:00 +00:00
desc = ist(stat_cols_px[ctx->field_num].desc);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, name))
continue;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!px)
px = proxies_list;
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
while (px) {
struct promex_label labels[PROMEX_MAX_LABELS-1] = {};
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[0].name = ist("proxy");
labels[0].value = ist2(px->id, strlen(px->id));
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
/* skip the disabled proxies, global frontend and non-networked ones */
MINOR: proxy: Introduce proxy flags to replace disabled bitfield This change is required to support TCP/HTTP rules in defaults sections. The 'disabled' bitfield in the proxy structure, used to know if a proxy is disabled or stopped, is replaced a generic bitfield named 'flags'. PR_DISABLED and PR_STOPPED flags are renamed to PR_FL_DISABLED and PR_FL_STOPPED respectively. In addition, everywhere there is a test to know if a proxy is disabled or stopped, there is now a bitwise AND operation on PR_FL_DISABLED and/or PR_FL_STOPPED flags.
2021-10-06 12:24:19 +00:00
if ((px->flags & PR_FL_DISABLED) || px->uuid <= 0 || !(px->cap & PR_CAP_BE))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto next_px;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!sv)
sv = px->srv;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
while (sv) {
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[1].name = ist("server");
labels[1].value = ist2(sv->id, strlen(sv->id));
MINOR: stats: use stricter naming stats/field/line Several unique names were used for different purposes under statistics implementation. This caused the code to be difficult to understand. * stat/stats name is removed when a more specific name could be used * restrict field usage to purely refer to <struct field> which represents a raw stat value. * use "line" naming to represent an array of <struct field>
2024-04-22 08:26:23 +00:00
if (!stats_fill_sv_line(px, sv, 0, stats, ST_I_PX_MAX, &(ctx->field_num)))
MINOR: contrib/prometheus-exporter: use fill_sv_stats for server dump use `stats_fill_sv_stats` when possible to avoid duplicating code. the following metrics have a change of behaviour: haproxy_server_limit_sessions haproxy_server_queue_limit haproxy_server_check_failures_total haproxy_server_check_up_down_total haproxy_server_downtime_seconds_total haproxy_server_current_throttle haproxy_server_idle_connections_limit depending on cases, if the limit was not configured or enabled, NaN is returned instead. It should not be an issue for users, even better than before as it provides more precise info. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-25 16:29:04 +00:00
return -1;
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
if ((ctx->flags & PROMEX_FL_NO_MAINT_SRV) && (sv->cur_admin & SRV_ADMF_MAINT))
MINOR: contrib/prometheus-exporter: Add a param to ignore servers in maintenance By passing the parameter "no-maint" in the query-string, it is now possible to ignore servers in maintenance. It means that the metrics for servers in this state will not be exported.
2019-11-19 13:18:24 +00:00
goto next_sv;
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
switch (ctx->field_num) {
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_STATUS:
MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for frontend/backend/servers states to labels values. the main motivation being I realised it is very difficult to make use of it without hard coded quirks on prometheus client side; especially because the main use is often to group by state, which is harder when the state is the value of the metric. in order to achieve that we iterate on the status metric to generate labels, and so as many metrics. this is the first step to resolve github issue #1029 A second step should address health check states. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:17 +00:00
state = promex_srv_status(sv);
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
for (; ctx->obj_state < PROMEX_SRV_STATE_COUNT; ctx->obj_state++) {
val = mkf_u32(FO_STATUS, state == ctx->obj_state);
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[2].name = ist("state");
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
labels[2].value = promex_srv_st[ctx->obj_state];
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for frontend/backend/servers states to labels values. the main motivation being I realised it is very difficult to make use of it without hard coded quirks on prometheus client side; especially because the main use is often to group by state, which is harder when the state is the value of the metric. in order to achieve that we iterate on the status metric to generate labels, and so as many metrics. this is the first step to resolve github issue #1029 A second step should address health check states. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:17 +00:00
goto full;
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->obj_state = 0;
MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for frontend/backend/servers states to labels values. the main motivation being I realised it is very difficult to make use of it without hard coded quirks on prometheus client side; especially because the main use is often to group by state, which is harder when the state is the value of the metric. in order to achieve that we iterate on the status metric to generate labels, and so as many metrics. this is the first step to resolve github issue #1029 A second step should address health check states. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-01-27 21:40:17 +00:00
goto next_sv;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_QTIME:
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
secs = (double)swrate_avg(sv->counters.q_time, TIME_STATS_SAMPLES) / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_AVG, secs);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_CTIME:
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
secs = (double)swrate_avg(sv->counters.c_time, TIME_STATS_SAMPLES) / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_AVG, secs);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_RTIME:
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
secs = (double)swrate_avg(sv->counters.d_time, TIME_STATS_SAMPLES) / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_AVG, secs);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_TTIME:
BUG/MINOR: contrib/prometheus-exporter: Return the time averages in seconds The metrics QTIME, CTIME, RTIME and TTIME are now returned in seconds using a float representation instead of in milliseconds. So these metrics are now consistent with their announced type and respect Prometheus naming conventions. This patch fixes the issue #288. It may be backported to 2.0. If so, the previous patch, introducing the support for float fields in stats is mantatory and should be backported first.
2019-09-24 14:35:19 +00:00
secs = (double)swrate_avg(sv->counters.t_time, TIME_STATS_SAMPLES) / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_AVG, secs);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_QT_MAX:
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
secs = (double)sv->counters.qtime_max / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_MAX, secs);
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_CT_MAX:
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
secs = (double)sv->counters.ctime_max / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_MAX, secs);
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_RT_MAX:
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
secs = (double)sv->counters.dtime_max / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_MAX, secs);
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_TT_MAX:
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
secs = (double)sv->counters.ttime_max / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_MAX, secs);
MINOR: contrib/prometheus-exporter: Report metrics about max times for sessions Now, for the sessions, the maximum times (queue, connect, response, total) are reported in addition of the averages over the last 1024 connections. These metrics are reported per backend and per server. Here are the metrics name : * haproxy_backend_max_queue_time_seconds * haproxy_backend_max_connect_time_seconds * haproxy_backend_max_response_time_seconds * haproxy_backend_max_total_time_seconds and * haproxy_server_max_queue_time_seconds * haproxy_server_max_connect_time_seconds * haproxy_server_max_response_time_seconds * haproxy_server_max_total_time_seconds This patch is related to #272.
2019-11-08 14:05:31 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_CHECK_STATUS:
MINOR: contrib/prometheus-exporter: Add heathcheck status/code in server metrics ST_F_CHECK_STATUS and ST_F_CHECK_CODE are now part of exported server metrics: * haproxy_server_check_status * haproxy_server_check_code The heathcheck status is an integer corresponding to HCHK_STATUS value.
2019-11-21 13:35:46 +00:00
if ((sv->check.state & (CHK_ST_ENABLED|CHK_ST_PAUSED)) != CHK_ST_ENABLED)
goto next_sv;
MAJOR: contrib/prometheus-exporter: move health check status to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for health check states to labels values. The diff is quite small thanks to the preparation work from Christopher to allow more flexibility in labels, see commit 5a2f938732126f43bbf4cea5482c01552b0d0314 ("MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones") this is a follow up of commit c6464591a365bfcf509b322bdaa4d608c9395d75 ("MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels"). The main goal being to be better aligned with prometheus use cases in terms of queries. More specifically to health checks, Pierre C. mentioned the possible quirks he had to put in place in order to make use of those metrics through prometheus: <aggregator_function> by(proxy, check_status) (count_values by(proxy, instance) ("check_status", haproxy_server_check_status)) I am perfectly aware this introduces a lot more metrics but I don't see how we can improve the usability without it. The main issue remains in the cardinality of the states which are > 20. Prometheus recommends to stay below a cardinality of 10 for a given metric but I consider our case very specific, because highly linked to the level of precision haproxy exposes. Even before this patch I saw several large production setup (a few hundreds of MB in output) which are making use of the scope parameter to simply ignore the server metrics, so that the scrapping can be faster, and memory consumed on client side not too high. So I believe we should eventually continue in that direction and offer more granularity of filtering of the output. That being said it is already possible to filter out the data on prometheus client side. this is related to github issue #1029 Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-01 12:11:51 +00:00
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
for (; ctx->obj_state < HCHK_STATUS_SIZE; ctx->obj_state++) {
if (get_check_status_result(ctx->obj_state) < CHK_RES_FAILED)
MAJOR: contrib/prometheus-exporter: move health check status to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for health check states to labels values. The diff is quite small thanks to the preparation work from Christopher to allow more flexibility in labels, see commit 5a2f938732126f43bbf4cea5482c01552b0d0314 ("MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones") this is a follow up of commit c6464591a365bfcf509b322bdaa4d608c9395d75 ("MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels"). The main goal being to be better aligned with prometheus use cases in terms of queries. More specifically to health checks, Pierre C. mentioned the possible quirks he had to put in place in order to make use of those metrics through prometheus: <aggregator_function> by(proxy, check_status) (count_values by(proxy, instance) ("check_status", haproxy_server_check_status)) I am perfectly aware this introduces a lot more metrics but I don't see how we can improve the usability without it. The main issue remains in the cardinality of the states which are > 20. Prometheus recommends to stay below a cardinality of 10 for a given metric but I consider our case very specific, because highly linked to the level of precision haproxy exposes. Even before this patch I saw several large production setup (a few hundreds of MB in output) which are making use of the scope parameter to simply ignore the server metrics, so that the scrapping can be faster, and memory consumed on client side not too high. So I believe we should eventually continue in that direction and offer more granularity of filtering of the output. That being said it is already possible to filter out the data on prometheus client side. this is related to github issue #1029 Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-01 12:11:51 +00:00
continue;
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
val = mkf_u32(FO_STATUS, sv->check.status == ctx->obj_state);
check_state = get_check_status_info(ctx->obj_state);
MAJOR: contrib/prometheus-exporter: move health check status to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for health check states to labels values. The diff is quite small thanks to the preparation work from Christopher to allow more flexibility in labels, see commit 5a2f938732126f43bbf4cea5482c01552b0d0314 ("MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones") this is a follow up of commit c6464591a365bfcf509b322bdaa4d608c9395d75 ("MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels"). The main goal being to be better aligned with prometheus use cases in terms of queries. More specifically to health checks, Pierre C. mentioned the possible quirks he had to put in place in order to make use of those metrics through prometheus: <aggregator_function> by(proxy, check_status) (count_values by(proxy, instance) ("check_status", haproxy_server_check_status)) I am perfectly aware this introduces a lot more metrics but I don't see how we can improve the usability without it. The main issue remains in the cardinality of the states which are > 20. Prometheus recommends to stay below a cardinality of 10 for a given metric but I consider our case very specific, because highly linked to the level of precision haproxy exposes. Even before this patch I saw several large production setup (a few hundreds of MB in output) which are making use of the scope parameter to simply ignore the server metrics, so that the scrapping can be faster, and memory consumed on client side not too high. So I believe we should eventually continue in that direction and offer more granularity of filtering of the output. That being said it is already possible to filter out the data on prometheus client side. this is related to github issue #1029 Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-01 12:11:51 +00:00
labels[2].name = ist("state");
CLEANUP: Apply ist.cocci This cleans up ist handling.
2021-09-15 11:58:44 +00:00
labels[2].value = ist(check_state);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
MAJOR: contrib/prometheus-exporter: move health check status to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for health check states to labels values. The diff is quite small thanks to the preparation work from Christopher to allow more flexibility in labels, see commit 5a2f938732126f43bbf4cea5482c01552b0d0314 ("MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones") this is a follow up of commit c6464591a365bfcf509b322bdaa4d608c9395d75 ("MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels"). The main goal being to be better aligned with prometheus use cases in terms of queries. More specifically to health checks, Pierre C. mentioned the possible quirks he had to put in place in order to make use of those metrics through prometheus: <aggregator_function> by(proxy, check_status) (count_values by(proxy, instance) ("check_status", haproxy_server_check_status)) I am perfectly aware this introduces a lot more metrics but I don't see how we can improve the usability without it. The main issue remains in the cardinality of the states which are > 20. Prometheus recommends to stay below a cardinality of 10 for a given metric but I consider our case very specific, because highly linked to the level of precision haproxy exposes. Even before this patch I saw several large production setup (a few hundreds of MB in output) which are making use of the scope parameter to simply ignore the server metrics, so that the scrapping can be faster, and memory consumed on client side not too high. So I believe we should eventually continue in that direction and offer more granularity of filtering of the output. That being said it is already possible to filter out the data on prometheus client side. this is related to github issue #1029 Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-01 12:11:51 +00:00
goto full;
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->obj_state = 0;
MAJOR: contrib/prometheus-exporter: move health check status to labels this patch is a breaking change between v2.3 and v2.4: we move from using gauge value for health check states to labels values. The diff is quite small thanks to the preparation work from Christopher to allow more flexibility in labels, see commit 5a2f938732126f43bbf4cea5482c01552b0d0314 ("MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones") this is a follow up of commit c6464591a365bfcf509b322bdaa4d608c9395d75 ("MAJOR: contrib/prometheus-exporter: move ftd/bkd/srv states to labels"). The main goal being to be better aligned with prometheus use cases in terms of queries. More specifically to health checks, Pierre C. mentioned the possible quirks he had to put in place in order to make use of those metrics through prometheus: <aggregator_function> by(proxy, check_status) (count_values by(proxy, instance) ("check_status", haproxy_server_check_status)) I am perfectly aware this introduces a lot more metrics but I don't see how we can improve the usability without it. The main issue remains in the cardinality of the states which are > 20. Prometheus recommends to stay below a cardinality of 10 for a given metric but I consider our case very specific, because highly linked to the level of precision haproxy exposes. Even before this patch I saw several large production setup (a few hundreds of MB in output) which are making use of the scope parameter to simply ignore the server metrics, so that the scrapping can be faster, and memory consumed on client side not too high. So I believe we should eventually continue in that direction and offer more granularity of filtering of the output. That being said it is already possible to filter out the data on prometheus client side. this is related to github issue #1029 Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-01 12:11:51 +00:00
goto next_sv;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_CHECK_CODE:
MINOR: contrib/prometheus-exporter: Add heathcheck status/code in server metrics ST_F_CHECK_STATUS and ST_F_CHECK_CODE are now part of exported server metrics: * haproxy_server_check_status * haproxy_server_check_code The heathcheck status is an integer corresponding to HCHK_STATUS value.
2019-11-21 13:35:46 +00:00
if ((sv->check.state & (CHK_ST_ENABLED|CHK_ST_PAUSED)) != CHK_ST_ENABLED)
goto next_sv;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_u32(FN_OUTPUT, (sv->check.status < HCHK_STATUS_L57DATA) ? 0 : sv->check.code);
MINOR: contrib/prometheus-exporter: Add heathcheck status/code in server metrics ST_F_CHECK_STATUS and ST_F_CHECK_CODE are now part of exported server metrics: * haproxy_server_check_status * haproxy_server_check_code The heathcheck status is an integer corresponding to HCHK_STATUS value.
2019-11-21 13:35:46 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_CHECK_DURATION:
MINOR: contrib/prometheus-exporter: Add the last heathcheck duration metric ST_F_CHECK_DURATION is now part of exported server metrics, named haproxy_server_check_duration_seconds and expressed in seconds. For a given server, this value is exported only if the healthcheck is finished (the status is greater or equal to HCHK_STATUS_CHECKED). This patch fixes the issue #519. It may be backported as fat as 2.0.
2020-02-27 15:12:07 +00:00
if (sv->check.status < HCHK_STATUS_CHECKED)
goto next_sv;
secs = (double)sv->check.duration / 1000.0;
MEDIUM: contrib/prometheus-exporter: Rework matrices defining Promex metrics The global and stats matrices are replaced by a simpler ones. Now we have only 2 arrays of prometheus metrics. Their flags are used to match on the entity type. This simplify a bit the metrics definition. For now, labels and descriptions are still outside of these arrays, because the labels must be reworked to be more dynamic and the descrptions must be replaced by stats ones as far as possible.
2021-01-20 14:20:53 +00:00
val = mkf_flt(FN_DURATION, secs);
MINOR: contrib/prometheus-exporter: Add the last heathcheck duration metric ST_F_CHECK_DURATION is now part of exported server metrics, named haproxy_server_check_duration_seconds and expressed in seconds. For a given server, this value is exported only if the healthcheck is finished (the status is greater or equal to HCHK_STATUS_CHECKED). This patch fixes the issue #519. It may be backported as fat as 2.0.
2020-02-27 15:12:07 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_REQ_TOT:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (px->mode != PR_MODE_HTTP) {
sv = NULL;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto next_px;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
}
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
break;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
case ST_I_PX_HRSP_1XX:
case ST_I_PX_HRSP_2XX:
case ST_I_PX_HRSP_3XX:
case ST_I_PX_HRSP_4XX:
case ST_I_PX_HRSP_5XX:
case ST_I_PX_HRSP_OTHER:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (px->mode != PR_MODE_HTTP) {
sv = NULL;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto next_px;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
}
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
if (ctx->field_num != ST_I_PX_HRSP_1XX)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~PROMEX_FL_METRIC_HDR;
MEDIUM: contrib/prometheus-exporter: Use dynamic labels instead of static ones Instead of using static labels for metrics, we now use an array of labels, filled for each metrics if necessary and passed to the dump function. This way, it is easier to extend the promex service. For now, there are at most 8 labels per metrics. This limit may be raised by changing PROMEX_MAX_LABELS value. And to ease labels addition, a label is defined as a key/value pair. The formatting is handled by the dump function. For the proxies and servers, the first entry of the array is always the proxy name. In addition, for the servers, the second entry is always the server name.
2021-01-28 10:24:17 +00:00
labels[2].name = ist("code");
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
labels[2].value = promex_hrsp_code[ctx->field_num - ST_I_PX_HRSP_1XX];
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
MINOR: contrib/prometheus-exporter: Add missing global and per-server metrics Following metrics are now exported by the prometheus exporter to reflect recent changes on HAProxy : * haproxy_process_failed_resolutions * haproxy_process_bytes_out_total * haproxy_process_spliced_bytes_out_total * haproxy_process_bytes_out_rate and * haproxy_server_unsafe_idle_connections_current * haproxy_server_safe_idle_connections_current * haproxy_server_used_connections_current * haproxy_server_need_connections_current
2020-07-10 13:39:39 +00:00
break;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
default:
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
val = stats[ctx->field_num];
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc,
promex_st_metrics[ctx->field_num].type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, &out, max))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto full;
MINOR: contrib/prometheus-exporter: Add a param to ignore servers in maintenance By passing the parameter "no-maint" in the query-string, it is now possible to ignore servers in maintenance. It means that the metrics for servers in this state will not be exported.
2019-11-19 13:18:24 +00:00
next_sv:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
sv = sv->next;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
next_px:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = px->next;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_METRIC_HDR;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
/* Skip extra counters */
if (!(ctx->flags & PROMEX_FL_EXTRA_COUNTERS))
goto end;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!mod) {
mod = LIST_NEXT(&stats_module_list[STATS_DOMAIN_PROXY], typeof(mod), list);
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
ctx->mod_field_num = 0;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
list_for_each_entry_from(mod, &stats_module_list[STATS_DOMAIN_PROXY], list) {
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
void *counters;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!(stats_px_get_cap(mod->domain_flags) & STATS_PX_CAP_SRV)) {
ctx->field_num += mod->stats_count;
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
ctx->mod_field_num = 0;
continue;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
for (;ctx->mod_field_num < mod->stats_count; ctx->mod_field_num++) {
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
name = ist2(mod->stats[ctx->mod_field_num].name, strlen(mod->stats[ctx->mod_field_num].name));
desc = ist2(mod->stats[ctx->mod_field_num].desc, strlen(mod->stats[ctx->mod_field_num].desc));
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, name))
continue;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!px)
px = proxies_list;
while (px) {
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
struct promex_label labels[PROMEX_MAX_LABELS-1] = {};
struct promex_metric metric;
labels[0].name = ist("proxy");
labels[0].value = ist2(px->id, strlen(px->id));
/* skip the disabled proxies, global frontend and non-networked ones */
if ((px->flags & PR_FL_DISABLED) || px->uuid <= 0 || !(px->cap & PR_CAP_BE))
goto next_px2;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
if (!sv)
sv = px->srv;
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
while (sv) {
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
labels[1].name = ist("server");
labels[1].value = ist2(sv->id, strlen(sv->id));
labels[2].name = ist("mod");
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
labels[2].value = ist2(mod->name, strlen(mod->name));
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
if ((ctx->flags & PROMEX_FL_NO_MAINT_SRV) && (sv->cur_admin & SRV_ADMF_MAINT))
goto next_sv2;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
counters = EXTRA_COUNTERS_GET(sv->extra_counters, mod);
if (!mod->fill_stats(counters, stats + ctx->field_num, &ctx->mod_field_num))
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
return -1;
val = stats[ctx->field_num + ctx->mod_field_num];
metric.type = ((val.type == FN_GAUGE) ? PROMEX_MT_GAUGE : PROMEX_MT_COUNTER);
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, name, desc, metric.type,
&val, labels, &out, max))
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
goto full;
next_sv2:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
sv = sv->next;
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
}
next_px2:
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = px->next;
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
}
ctx->flags |= PROMEX_FL_METRIC_HDR;
}
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
ctx->field_num += mod->stats_count;
MEDIUM: promex: Dump servers extra counters if requested Dump extra counter for the servers. We loop on stats modules and dumped all concerned counters. The module name is reported with the "mod" label.
2024-01-29 15:41:56 +00:00
ctx->mod_field_num = 0;
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
px = NULL;
sv = NULL;
mod = NULL;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
end:
BUG/MINOR: contrib/prometheus-exporter: Don't try to add empty data blocks When the response buffer is full and nothing more can be inserted, it is important to not try to insert an empty data block. Otherwise, when the function channel_add_input() is called, the flag CF_READ_PARTIAL is set on the response channel while nothing was read and the stream is uselessly woken up. Finally, we have loop while the response buffer is full. This patch must be backported to 2.0.
2019-07-04 08:03:28 +00:00
if (out.len) {
if (!htx_add_data_atonce(htx, out))
return -1; /* Unexpected and unrecoverable error */
channel_add_input(chn, out.len);
}
BUG/MAJOR: promex: fix crash on deleted server Promex applet is used to dump many metrics. Some of them are related to a server instance. The applet can be interrupted in the middle of a dump, for example waiting for output buffer space. In this case, its context is save to resume dump on the correct instance. A crash can occur if dump is interrupted during servers loop. If the server instance is deleted during two scheduling of the promex applet, its context will still referenced the deleted server on resume. To fix this, use server refcount to prevent its deletion during parsing. No backport is needed, despite all stable releases being affected. This is because promex applet context has been recently rewritten to use generic pointers. As such, a specific commit will be applied for earlier releases.
2024-02-22 13:16:37 +00:00
/* Decrement server refcount if it was saved through ctx.p[1]. */
srv_drop(ctx->p[1]);
if (sv)
srv_take(sv);
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
/* Save pointers (0=current proxy, 1=current server, 2=current stats module) of the current context */
ctx->p[0] = px;
ctx->p[1] = sv;
ctx->p[2] = mod;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return ret;
full:
ret = 0;
goto end;
}
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
/* Dump metrics of module <mod>. It returns 1 on success, 0 if <out> is full and
* -1 on error. */
static int promex_dump_module_metrics(struct appctx *appctx, struct promex_module *mod,
struct ist *out, size_t max)
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
{
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
struct ist prefix = { .ptr = (char[PROMEX_MAX_NAME_LEN]){ 0 }, .len = 0 };
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
struct promex_ctx *ctx = appctx->svcctx;
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
int ret = 1;
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
istcat(&prefix, ist("haproxy_"), PROMEX_MAX_NAME_LEN);
istcat(&prefix, mod->name, PROMEX_MAX_NAME_LEN);
istcat(&prefix, ist("_"), PROMEX_MAX_NAME_LEN);
if (!ctx->p[1] && mod->start_metrics_dump) {
ctx->p[1] = mod->start_metrics_dump();
if (!ctx->p[1])
goto end;
}
for (; ctx->mod_field_num < mod->nb_metrics; ctx->mod_field_num++) {
struct promex_metric metric;
struct ist desc;
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
ret = mod->metric_info(ctx->mod_field_num, &metric, &desc);
if (!ret)
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
continue;
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
if (ret < 0)
goto error;
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
if (promex_filter_metric(appctx, prefix, metric.n))
continue;
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
if (!ctx->p[2])
ctx->p[2] = mod->start_ts(ctx->p[1], ctx->mod_field_num);
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
while (ctx->p[2]) {
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
struct promex_label labels[PROMEX_MAX_LABELS - 1] = {};
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
struct field val;
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
ret = mod->fill_ts(ctx->p[1], ctx->p[2], ctx->mod_field_num, labels, &val);
if (!ret)
continue;
if (ret < 0)
goto error;
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
MINOR: promex: Always pass the final name and description to promex_dmp_ts() It is easier this way, especially for promex modules. And because name and description are now explicitly passed to this function, there is no reason to still pass the metric, its type is enough. The function is easier to read this way.
2024-01-31 10:54:58 +00:00
if (!promex_dump_ts(appctx, prefix, metric.n, desc, metric.type,
MINOR: promex: Rename dump functions to use the right wording In Prometheus, a time series a stream of timestamped values belonging to the same metric and the same set of labeled dimensions. Thus the exporter dump time-series and not metrics. Thus, promex_dump_metric(), promex_dump_metric_header() and promex_metric_to_str() functions were renamed to replace "metric"
2024-01-31 09:42:10 +00:00
&val, labels, out, max))
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
goto full;
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
next:
ctx->p[2] = mod->next_ts(ctx->p[1], ctx->p[2], ctx->mod_field_num);
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_METRIC_HDR;
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
}
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
ret = 1;
end:
if (ctx->p[1] && mod->stop_metrics_dump)
mod->stop_metrics_dump(ctx->p[1]);
ctx->p[1] = NULL;
ctx->p[2] = NULL;
return ret;
full:
return 0;
error:
ret = -1;
goto end;
}
/* Dump metrics of referenced modules. It returns 1 on success, 0 if <htx> is
* full and -1 in case of any error. */
static int promex_dump_ref_modules_metrics(struct appctx *appctx, struct htx *htx)
{
struct promex_ctx *ctx = appctx->svcctx;
struct promex_module_ref *ref = ctx->p[0];
struct channel *chn = sc_ic(appctx_sc(appctx));
struct ist out = ist2(trash.area, 0);
size_t max = htx_get_max_blksz(htx, channel_htx_recv_max(chn, htx));
int ret = 1;
if (!ref) {
ref = LIST_NEXT(&ctx->modules, typeof(ref), list);
ctx->mod_field_num = 0;
}
list_for_each_entry_from(ref, &ctx->modules, list) {
ret = promex_dump_module_metrics(appctx, ref->mod, &out, max);
if (ret <= 0) {
if (ret == -1)
return -1;
goto full;
}
ctx->mod_field_num = 0;
}
ref = NULL;
end:
if (out.len) {
if (!htx_add_data_atonce(htx, out))
return -1; /* Unexpected and unrecoverable error */
channel_add_input(chn, out.len);
}
ctx->p[0] = ref;
return ret;
full:
ret = 0;
goto end;
}
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
/* Dump metrics of all registered modules. It returns 1 on success, 0 if <htx> is
* full and -1 in case of any error. */
static int promex_dump_all_modules_metrics(struct appctx *appctx, struct htx *htx)
{
struct promex_ctx *ctx = appctx->svcctx;
struct promex_module *mod = ctx->p[0];
struct channel *chn = sc_ic(appctx_sc(appctx));
struct ist out = ist2(trash.area, 0);
size_t max = htx_get_max_blksz(htx, channel_htx_recv_max(chn, htx));
int ret = 1;
if (!mod) {
mod = LIST_NEXT(&promex_module_list, typeof(mod), list);
ctx->mod_field_num = 0;
}
list_for_each_entry_from(mod, &promex_module_list, list) {
ret = promex_dump_module_metrics(appctx, mod, &out, max);
if (ret <= 0) {
if (ret == -1)
return -1;
goto full;
}
ctx->mod_field_num = 0;
}
mod = NULL;
MEDIUM: promex: Simplify the context using generic pointers for restart points Instead of using typed pointers to save the restart points we know use generic pointers. 4 pointers can be saved now. This replaces the 5 typed pointers used before. So, we save 8-bytes but it is also more generic and this will be used by the promex modules.
2024-01-30 13:36:07 +00:00
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
end:
if (out.len) {
if (!htx_add_data_atonce(htx, out))
return -1; /* Unexpected and unrecoverable error */
channel_add_input(chn, out.len);
}
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
ctx->p[0] = mod;
MEDIUM: contrib/prometheus-exporter: export base stick table stats I saw some people falling back to unix socket to collect some data they could not find in prometheus exporter. One of them is base info from stick tables (used/size). I do not plan to extend it more for now; keys are quite a mess to handle. This should resolve github issue #1008. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-07 19:42:38 +00:00
return ret;
full:
ret = 0;
goto end;
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
/* Dump all metrics (global, frontends, backends and servers) depending on the
* dumper state (appctx->st1). It returns 1 on success, 0 if <htx> is full and
MINOR: stats: hide px/sv/li fields in applet struct Use an opaque pointer to store proxy instance. Regroup server/listener as a single opaque pointer. This has the benefit to render the structure more evolutive to support statistics on other types of objects in the future. This patch is needed to extend stat support for components other than proxies objects. The prometheus module has been adapted for these changes.
2020-10-02 16:32:02 +00:00
* -1 in case of any error.
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
* Uses <appctx.ctx.stats.px> as a pointer to the current proxy and <sv>/<li>
* as pointers to the current server/listener respectively.
*/
CLEANUP: promex: rename all occurrences of stconn "cs" to "sc" Function arguments and local variables called "cs" were renamed to "sc" to avoid future confusion.
2022-05-27 08:16:15 +00:00
static int promex_dump_metrics(struct appctx *appctx, struct stconn *sc, struct htx *htx)
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
{
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
struct promex_ctx *ctx = appctx->svcctx;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
int ret;
switch (appctx->st1) {
case PROMEX_DUMPER_INIT:
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= (PROMEX_FL_METRIC_HDR|PROMEX_FL_INFO_METRIC);
ctx->obj_state = 0;
MINOR: stats: rename info stats Info are used to expose haproxy global metrics. It is similar to proxy statistics and any other module. As such, rename info indexes using SI_I_INF_* prefix. Also info variable is renamed stat_line_info. Thanks to this, naming is now consistent between info and other statistics. It will help to integrate it as a "global" statistics module.
2024-04-22 07:41:15 +00:00
ctx->field_num = ST_I_INF_NAME;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
appctx->st1 = PROMEX_DUMPER_GLOBAL;
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
case PROMEX_DUMPER_GLOBAL:
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
if (ctx->flags & PROMEX_FL_SCOPE_GLOBAL) {
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
ret = promex_dump_global_metrics(appctx, htx);
if (ret <= 0) {
if (ret == -1)
goto error;
goto full;
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~PROMEX_FL_INFO_METRIC;
ctx->flags |= (PROMEX_FL_METRIC_HDR|PROMEX_FL_FRONT_METRIC);
ctx->obj_state = 0;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
ctx->field_num = ST_I_PX_PXNAME;
MINOR: promex: Add info in the promex context to dump extra counters The context of the promex applet was extended to support the dump of extra counters. These counters are not dumped yet, but info to interrupt and restart the dump are required. The stats module and the relative field number for this module can now be saved. In addition support for "extra-counters" parameter was added on the query-string to dump these counters. Otherwise, no extra-counters are dumped.
2024-01-29 15:40:41 +00:00
ctx->mod_field_num = 0;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
appctx->st1 = PROMEX_DUMPER_FRONT;
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
case PROMEX_DUMPER_FRONT:
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
if (ctx->flags & PROMEX_FL_SCOPE_FRONT) {
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
ret = promex_dump_front_metrics(appctx, htx);
if (ret <= 0) {
if (ret == -1)
goto error;
goto full;
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~PROMEX_FL_FRONT_METRIC;
ctx->flags |= (PROMEX_FL_METRIC_HDR|PROMEX_FL_LI_METRIC);
ctx->obj_state = 0;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
ctx->field_num = ST_I_PX_PXNAME;
MINOR: promex: Add info in the promex context to dump extra counters The context of the promex applet was extended to support the dump of extra counters. These counters are not dumped yet, but info to interrupt and restart the dump are required. The stats module and the relative field number for this module can now be saved. In addition support for "extra-counters" parameter was added on the query-string to dump these counters. Otherwise, no extra-counters are dumped.
2024-01-29 15:40:41 +00:00
ctx->mod_field_num = 0;
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
appctx->st1 = PROMEX_DUMPER_LI;
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
case PROMEX_DUMPER_LI:
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
if (ctx->flags & PROMEX_FL_SCOPE_LI) {
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
ret = promex_dump_listener_metrics(appctx, htx);
if (ret <= 0) {
if (ret == -1)
goto error;
goto full;
}
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~PROMEX_FL_LI_METRIC;
ctx->flags |= (PROMEX_FL_METRIC_HDR|PROMEX_FL_BACK_METRIC);
ctx->obj_state = 0;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
ctx->field_num = ST_I_PX_PXNAME;
MINOR: promex: Add info in the promex context to dump extra counters The context of the promex applet was extended to support the dump of extra counters. These counters are not dumped yet, but info to interrupt and restart the dump are required. The stats module and the relative field number for this module can now be saved. In addition support for "extra-counters" parameter was added on the query-string to dump these counters. Otherwise, no extra-counters are dumped.
2024-01-29 15:40:41 +00:00
ctx->mod_field_num = 0;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
appctx->st1 = PROMEX_DUMPER_BACK;
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
case PROMEX_DUMPER_BACK:
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
if (ctx->flags & PROMEX_FL_SCOPE_BACK) {
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
ret = promex_dump_back_metrics(appctx, htx);
if (ret <= 0) {
if (ret == -1)
goto error;
goto full;
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~PROMEX_FL_BACK_METRIC;
ctx->flags |= (PROMEX_FL_METRIC_HDR|PROMEX_FL_SRV_METRIC);
ctx->obj_state = 0;
MINOR: stats: rename proxy stats This commit is the first one of a serie which adjust naming convention for stats module. The objective is to remove ambiguity and better reflect how stats are implemented, especially since the introduction of stats module. This patch renames elements related to proxies statistics. One of the main change is to rename ST_F_* statistics indexes prefix with the new name ST_I_PX_*. This remove the reference to field which represents another concept in the stats module. In the same vein, global stat_fields variable is renamed metrics_px.
2024-04-19 16:03:45 +00:00
ctx->field_num = ST_I_PX_PXNAME;
MINOR: promex: Add info in the promex context to dump extra counters The context of the promex applet was extended to support the dump of extra counters. These counters are not dumped yet, but info to interrupt and restart the dump are required. The stats module and the relative field number for this module can now be saved. In addition support for "extra-counters" parameter was added on the query-string to dump these counters. Otherwise, no extra-counters are dumped.
2024-01-29 15:40:41 +00:00
ctx->mod_field_num = 0;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
appctx->st1 = PROMEX_DUMPER_SRV;
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
case PROMEX_DUMPER_SRV:
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
if (ctx->flags & PROMEX_FL_SCOPE_SERVER) {
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
ret = promex_dump_srv_metrics(appctx, htx);
if (ret <= 0) {
if (ret == -1)
goto error;
goto full;
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~(PROMEX_FL_METRIC_HDR|PROMEX_FL_SRV_METRIC);
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
ctx->flags |= (PROMEX_FL_METRIC_HDR|PROMEX_FL_MODULE_METRIC);
ctx->field_num = 0;
ctx->mod_field_num = 0;
appctx->st1 = PROMEX_DUMPER_MODULES;
__fallthrough;
case PROMEX_DUMPER_MODULES:
if (ctx->flags & PROMEX_FL_SCOPE_MODULE) {
if (LIST_ISEMPTY(&ctx->modules))
ret = promex_dump_all_modules_metrics(appctx, htx);
else
ret = promex_dump_ref_modules_metrics(appctx, htx);
if (ret <= 0) {
if (ret == -1)
goto error;
goto full;
}
}
ctx->flags &= ~(PROMEX_FL_METRIC_HDR|PROMEX_FL_MODULE_METRIC);
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
ctx->field_num = 0;
MINOR: promex: Add info in the promex context to dump extra counters The context of the promex applet was extended to support the dump of extra counters. These counters are not dumped yet, but info to interrupt and restart the dump are required. The stats module and the relative field number for this module can now be saved. In addition support for "extra-counters" parameter was added on the query-string to dump these counters. Otherwise, no extra-counters are dumped.
2024-01-29 15:40:41 +00:00
ctx->mod_field_num = 0;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
appctx->st1 = PROMEX_DUMPER_DONE;
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
case PROMEX_DUMPER_DONE:
default:
break;
}
return 1;
full:
MEDIUM: tree-wide: Change sc API to specify required free space to progress sc_need_room() now takes the required free space to receive more data as parameter. All calls to this function are updated accordingly. For now, this value is set but not used. When we are waiting for a buffer, 0 is used. So we expect to be unblocked ASAP. However this must be reviewed because SC_FL_NEED_BUF is probably enough in this case and this flag is already set if the input buffer allocation fails.
2023-05-05 09:28:45 +00:00
sc_need_room(sc, channel_htx_recv_max(sc_ic(appctx_sc(appctx)), htx) + 1);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return 0;
error:
/* unrecoverable error */
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags = 0;
CLEANUP: promex: stop using appctx->st2 Now that we have out own context, there's no need to use the cryptic st2 struct member. It's solely used to carry a field number here, so let's add this into the context.
2022-05-03 16:05:23 +00:00
ctx->field_num = 0;
MINOR: promex: Add info in the promex context to dump extra counters The context of the promex applet was extended to support the dump of extra counters. These counters are not dumped yet, but info to interrupt and restart the dump are required. The stats module and the relative field number for this module can now be saved. In addition support for "extra-counters" parameter was added on the query-string to dump these counters. Otherwise, no extra-counters are dumped.
2024-01-29 15:40:41 +00:00
ctx->mod_field_num = 0;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
appctx->st1 = PROMEX_DUMPER_DONE;
return -1;
}
CLEANUP: assorted typo fixes in the code and comments This is fifth iteration of typo fixes
2020-03-23 17:28:40 +00:00
/* Parse the query string of request URI to filter the metrics. It returns 1 on
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
* success and -1 on error. */
CLEANUP: promex: rename all occurrences of stconn "cs" to "sc" Function arguments and local variables called "cs" were renamed to "sc" to avoid future confusion.
2022-05-27 08:16:15 +00:00
static int promex_parse_uri(struct appctx *appctx, struct stconn *sc)
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
{
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
struct promex_ctx *ctx = appctx->svcctx;
CLEANUP: promex: rename all occurrences of stconn "cs" to "sc" Function arguments and local variables called "cs" were renamed to "sc" to avoid future confusion.
2022-05-27 08:16:15 +00:00
struct channel *req = sc_oc(sc);
struct channel *res = sc_ic(sc);
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
struct htx *req_htx, *res_htx;
struct htx_sl *sl;
BUG/MINOR: contrib/prometheus-exporter: decode parameter and value only we were decoding all substring and then parsing; this could lead to consider & and = in decoding result as delimiters where it should not. this patch reverses the order by first parsing and then decoding each key and value separately. we also stop parsing after number sign (#). This patch should be backported to 2.1 and 2.0 Signed-off-by: William Dauchy <w.dauchy@criteo.com>
2019-11-26 11:56:26 +00:00
char *p, *key, *value;
const char *end;
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
struct buffer *err;
int default_scopes = PROMEX_FL_SCOPE_ALL;
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
int default_metrics_filter = PROMEX_FL_INC_METRIC_BY_DEFAULT;
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
int len;
/* Get the query-string */
req_htx = htxbuf(&req->buf);
sl = http_get_stline(req_htx);
if (!sl)
goto error;
p = http_find_param_list(HTX_SL_REQ_UPTR(sl), HTX_SL_REQ_ULEN(sl), '?');
if (!p)
goto end;
end = HTX_SL_REQ_UPTR(sl) + HTX_SL_REQ_ULEN(sl);
BUG/MINOR: contrib/prometheus-exporter: decode parameter and value only we were decoding all substring and then parsing; this could lead to consider & and = in decoding result as delimiters where it should not. this patch reverses the order by first parsing and then decoding each key and value separately. we also stop parsing after number sign (#). This patch should be backported to 2.1 and 2.0 Signed-off-by: William Dauchy <w.dauchy@criteo.com>
2019-11-26 11:56:26 +00:00
/* copy the query-string */
len = end - p;
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
chunk_reset(&trash);
memcpy(trash.area, p, len);
trash.area[len] = 0;
p = trash.area;
BUG/MINOR: contrib/prometheus-exporter: decode parameter and value only we were decoding all substring and then parsing; this could lead to consider & and = in decoding result as delimiters where it should not. this patch reverses the order by first parsing and then decoding each key and value separately. we also stop parsing after number sign (#). This patch should be backported to 2.1 and 2.0 Signed-off-by: William Dauchy <w.dauchy@criteo.com>
2019-11-26 11:56:26 +00:00
end = trash.area + len;
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
/* Parse the query-string */
BUG/MINOR: contrib/prometheus-exporter: decode parameter and value only we were decoding all substring and then parsing; this could lead to consider & and = in decoding result as delimiters where it should not. this patch reverses the order by first parsing and then decoding each key and value separately. we also stop parsing after number sign (#). This patch should be backported to 2.1 and 2.0 Signed-off-by: William Dauchy <w.dauchy@criteo.com>
2019-11-26 11:56:26 +00:00
while (p < end && *p && *p != '#') {
value = NULL;
/* decode parameter name */
key = p;
while (p < end && *p != '=' && *p != '&' && *p != '#')
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
++p;
BUG/MINOR: contrib/prometheus-exporter: decode parameter and value only we were decoding all substring and then parsing; this could lead to consider & and = in decoding result as delimiters where it should not. this patch reverses the order by first parsing and then decoding each key and value separately. we also stop parsing after number sign (#). This patch should be backported to 2.1 and 2.0 Signed-off-by: William Dauchy <w.dauchy@criteo.com>
2019-11-26 11:56:26 +00:00
/* found a value */
if (*p == '=') {
*(p++) = 0;
value = p;
}
else if (*p == '&')
*(p++) = 0;
else if (*p == '#')
*p = 0;
BUG/MINOR: http: make url_decode() optionally convert '+' to SP The url_decode() function used by the url_dec converter and a few other call points is ambiguous on its processing of the '+' character which itself isn't stable in the spec. This one belongs to the reserved characters for the query string but not for the path nor the scheme, in which it must be left as-is. It's only in argument strings that follow the application/x-www-form-urlencoded encoding that it must be turned into a space, that is, in query strings and POST arguments. The problem is that the function is used to process full URLs and paths in various configs, and to process query strings from the stats page for example. This patch updates the function to differentiate the situation where it's parsing a path and a query string. A new argument indicates if a query string should be assumed, otherwise it's only assumed after seeing a question mark. The various locations in the code making use of this function were updated to take care of this (most call places were using it to decode POST arguments). The url_dec converter is usually called on path or url samples, so it needs to remain compatible with this and will default to parsing a path and turning the '+' to a space only after a question mark. However in situations where it would explicitly be extracted from a POST or a query string, it now becomes possible to enforce the decoding by passing a non-null value in argument. It seems to be what was reported in issue #585. This fix may be backported to older stable releases.
2020-04-23 15:54:47 +00:00
len = url_decode(key, 1);
BUG/MINOR: contrib/prometheus-exporter: decode parameter and value only we were decoding all substring and then parsing; this could lead to consider & and = in decoding result as delimiters where it should not. this patch reverses the order by first parsing and then decoding each key and value separately. we also stop parsing after number sign (#). This patch should be backported to 2.1 and 2.0 Signed-off-by: William Dauchy <w.dauchy@criteo.com>
2019-11-26 11:56:26 +00:00
if (len == -1)
goto error;
/* decode value */
if (value) {
while (p < end && *p != '=' && *p != '&' && *p != '#')
++p;
if (*p == '=')
goto error;
if (*p == '&')
*(p++) = 0;
else if (*p == '#')
*p = 0;
BUG/MINOR: http: make url_decode() optionally convert '+' to SP The url_decode() function used by the url_dec converter and a few other call points is ambiguous on its processing of the '+' character which itself isn't stable in the spec. This one belongs to the reserved characters for the query string but not for the path nor the scheme, in which it must be left as-is. It's only in argument strings that follow the application/x-www-form-urlencoded encoding that it must be turned into a space, that is, in query strings and POST arguments. The problem is that the function is used to process full URLs and paths in various configs, and to process query strings from the stats page for example. This patch updates the function to differentiate the situation where it's parsing a path and a query string. A new argument indicates if a query string should be assumed, otherwise it's only assumed after seeing a question mark. The various locations in the code making use of this function were updated to take care of this (most call places were using it to decode POST arguments). The url_dec converter is usually called on path or url samples, so it needs to remain compatible with this and will default to parsing a path and turning the '+' to a space only after a question mark. However in situations where it would explicitly be extracted from a POST or a query string, it now becomes possible to enforce the decoding by passing a non-null value in argument. It seems to be what was reported in issue #585. This fix may be backported to older stable releases.
2020-04-23 15:54:47 +00:00
len = url_decode(value, 1);
BUG/MINOR: contrib/prometheus-exporter: decode parameter and value only we were decoding all substring and then parsing; this could lead to consider & and = in decoding result as delimiters where it should not. this patch reverses the order by first parsing and then decoding each key and value separately. we also stop parsing after number sign (#). This patch should be backported to 2.1 and 2.0 Signed-off-by: William Dauchy <w.dauchy@criteo.com>
2019-11-26 11:56:26 +00:00
if (len == -1)
goto error;
}
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
CLEANUP: Apply the coccinelle patch for `XXXcmp()` on contrib/ Compare the various `cmp()` functions against zero.
2021-01-02 21:31:55 +00:00
if (strcmp(key, "scope") == 0) {
BUG/MINOR: contrib/prometheus-exporter: decode parameter and value only we were decoding all substring and then parsing; this could lead to consider & and = in decoding result as delimiters where it should not. this patch reverses the order by first parsing and then decoding each key and value separately. we also stop parsing after number sign (#). This patch should be backported to 2.1 and 2.0 Signed-off-by: William Dauchy <w.dauchy@criteo.com>
2019-11-26 11:56:26 +00:00
default_scopes = 0; /* at least a scope defined, unset default scopes */
if (!value)
goto error;
else if (*value == 0)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags &= ~PROMEX_FL_SCOPE_ALL;
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
else if (*value == '*' && *(value+1) == 0)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_SCOPE_ALL;
CLEANUP: Apply the coccinelle patch for `XXXcmp()` on contrib/ Compare the various `cmp()` functions against zero.
2021-01-02 21:31:55 +00:00
else if (strcmp(value, "global") == 0)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_SCOPE_GLOBAL;
CLEANUP: Apply the coccinelle patch for `XXXcmp()` on contrib/ Compare the various `cmp()` functions against zero.
2021-01-02 21:31:55 +00:00
else if (strcmp(value, "server") == 0)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_SCOPE_SERVER;
CLEANUP: Apply the coccinelle patch for `XXXcmp()` on contrib/ Compare the various `cmp()` functions against zero.
2021-01-02 21:31:55 +00:00
else if (strcmp(value, "backend") == 0)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_SCOPE_BACK;
CLEANUP: Apply the coccinelle patch for `XXXcmp()` on contrib/ Compare the various `cmp()` functions against zero.
2021-01-02 21:31:55 +00:00
else if (strcmp(value, "frontend") == 0)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_SCOPE_FRONT;
MEDIUM: contrib/prometheus-exporter: add listen stats this was a missing piece for a while now even though it was planned. This patch adds listen stats. Nothing in particular but we make use of the status helper previously added. `promex_st_metrics` diff also looks scary, but I had to realign all lines. Signed-off-by: William Dauchy <wdauchy@gmail.com>
2021-02-14 22:22:56 +00:00
else if (strcmp(value, "listener") == 0)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_SCOPE_LI;
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
else {
struct promex_module *mod;
struct promex_module_ref *ref;
list_for_each_entry(mod, &promex_module_list, list) {
if (strncmp(value, istptr(mod->name), istlen(mod->name)) == 0) {
ref = pool_alloc(pool_head_promex_mod_ref);
if (!ref)
goto internal_error;
ctx->flags |= PROMEX_FL_SCOPE_MODULE;
ref->mod = mod;
LIST_APPEND(&ctx->modules, &ref->list);
break;
}
}
if (!(ctx->flags & PROMEX_FL_SCOPE_MODULE))
goto error;
}
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
}
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
else if (strcmp(key, "metrics") == 0) {
struct ist args;
if (!value)
goto error;
for (args = ist(value); istlen(args); args = istadv(istfind(args, ','), 1)) {
struct eb32_node *node;
struct promex_metric_filter *flt;
struct ist m = iststop(args, ',');
unsigned int hash;
int exclude = 0;
if (!istlen(m))
continue;
if (*istptr(m) == '-') {
m = istnext(m);
if (!istlen(m))
continue;
exclude = 1;
}
else
default_metrics_filter &= ~PROMEX_FL_INC_METRIC_BY_DEFAULT;
hash = XXH32(istptr(m), istlen(m), 0);
node = eb32_lookup(&ctx->filters, hash);
if (node) {
flt = container_of(node, typeof(*flt), node);
flt->exclude = exclude;
continue;
}
flt = pool_alloc(pool_head_promex_metric_flt);
if (!flt)
goto internal_error;
flt->node.key = hash;
flt->exclude = exclude;
eb32_insert(&ctx->filters, &flt->node);
}
}
MINOR: promex: Add info in the promex context to dump extra counters The context of the promex applet was extended to support the dump of extra counters. These counters are not dumped yet, but info to interrupt and restart the dump are required. The stats module and the relative field number for this module can now be saved. In addition support for "extra-counters" parameter was added on the query-string to dump these counters. Otherwise, no extra-counters are dumped.
2024-01-29 15:40:41 +00:00
else if (strcmp(key, "extra-counters") == 0) {
ctx->flags |= PROMEX_FL_EXTRA_COUNTERS;
}
CLEANUP: Apply the coccinelle patch for `XXXcmp()` on contrib/ Compare the various `cmp()` functions against zero.
2021-01-02 21:31:55 +00:00
else if (strcmp(key, "no-maint") == 0)
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
ctx->flags |= PROMEX_FL_NO_MAINT_SRV;
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
}
end:
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
ctx->flags |= (default_scopes | default_metrics_filter);
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
return 1;
error:
err = &http_err_chunks[HTTP_ERR_400];
channel_erase(res);
res->buf.data = b_data(err);
memcpy(res->buf.area, b_head(err), b_data(err));
res_htx = htx_from_buf(&res->buf);
channel_add_input(res, res_htx->data);
return -1;
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
internal_error:
err = &http_err_chunks[HTTP_ERR_400];
channel_erase(res);
res->buf.data = b_data(err);
memcpy(res->buf.area, b_head(err), b_data(err));
res_htx = htx_from_buf(&res->buf);
channel_add_input(res, res_htx->data);
return -1;
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
/* Send HTTP headers of the response. It returns 1 on success and 0 if <htx> is
* full. */
CLEANUP: promex: rename all occurrences of stconn "cs" to "sc" Function arguments and local variables called "cs" were renamed to "sc" to avoid future confusion.
2022-05-27 08:16:15 +00:00
static int promex_send_headers(struct appctx *appctx, struct stconn *sc, struct htx *htx)
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
{
CLEANUP: promex: rename all occurrences of stconn "cs" to "sc" Function arguments and local variables called "cs" were renamed to "sc" to avoid future confusion.
2022-05-27 08:16:15 +00:00
struct channel *chn = sc_ic(sc);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
struct htx_sl *sl;
unsigned int flags;
flags = (HTX_SL_F_IS_RESP|HTX_SL_F_VER_11|HTX_SL_F_XFER_ENC|HTX_SL_F_XFER_LEN|HTX_SL_F_CHNK);
sl = htx_add_stline(htx, HTX_BLK_RES_SL, flags, ist("HTTP/1.1"), ist("200"), ist("OK"));
if (!sl)
goto full;
sl->info.res.status = 200;
if (!htx_add_header(htx, ist("Cache-Control"), ist("no-cache")) ||
!htx_add_header(htx, ist("Content-Type"), ist("text/plain; version=0.0.4")) ||
!htx_add_header(htx, ist("Transfer-Encoding"), ist("chunked")) ||
!htx_add_endof(htx, HTX_BLK_EOH))
goto full;
BUG/MINOR: contrib/prometheus-exporter: Fix applet accordingly to recent changes Since the flag EOI was added on channels, some hidden bugs in the prometheus exporter now leads to error. the visible effect is that responses are truncated. So first of all, channel_add_input() must be called when the response headers and the EOM block are added. To be sure to correctly update the response channel (especially to_forward value). Then the request must really be fully consumed. And finally, the return clause in the switch has been replaced by a break. It was totally wrong to skip the end of the function in the states PROMEX_DONE and PROMEX_ERROR. (Note that PROMEX_ERROR was never used, so it was replaced by PROMEX_END just to ease reading the code). No need to backport this patch, the Prometheus exporter does not exist in early versions.
2019-03-27 14:48:53 +00:00
channel_add_input(chn, htx->data);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return 1;
full:
htx_reset(htx);
MEDIUM: tree-wide: Change sc API to specify required free space to progress sc_need_room() now takes the required free space to receive more data as parameter. All calls to this function are updated accordingly. For now, this value is set but not used. When we are waiting for a buffer, 0 is used. So we expect to be unblocked ASAP. However this must be reviewed because SC_FL_NEED_BUF is probably enough in this case and this flag is already set if the input buffer allocation fails.
2023-05-05 09:28:45 +00:00
sc_need_room(sc, 0);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return 0;
}
/* The function returns 1 if the initialisation is complete, 0 if
* an errors occurs and -1 if more data are required for initializing
* the applet.
*/
MINOR: applet: Make .init callback more generic For now there is no much change. Only the appctx is passed as argument when the .init callback function is called. And it is not possible to yield at this stage. It is not a problem because the feature is not used. Only the lua defines this callback function for the lua TCP/HTTP services. The idea is to be able to use it for all applets to initialize the appctx context.
2022-01-13 15:01:35 +00:00
static int promex_appctx_init(struct appctx *appctx)
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
{
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
struct promex_ctx *ctx;
CLEANUP: promex: make the applet use its own context The prometheus applet used to map to the stats context since it was not convenient to have one's own context, and to reuse the fields with its own values and enums. The obj1 pointer was used both for proxies and stick-tables; obj2 was used both for servers and listeners. This change makes use of the generic command context allocation so that the there's now a properly typed context for prometheus, defined in the code itself and independent on the stats or appctx ones. For clarity, the types are correctly set and there's one proxy, one table, one server and one listener. Some could be compacted using unions but that's not necessary since the context is reasonably compact. The stats' st_code field was used as the object state so the new field name is obj_state. An attempt was made to change the types to const for what us only visited but some calls pass through the stats code to retrieve info and that code uses non-const variables due to internal API limitations (read_freq_ctr() being used and requiring variable). That could change in the future, though.
2022-05-03 15:33:00 +00:00
applet_reserve_svcctx(appctx, sizeof(struct promex_ctx));
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
ctx = appctx->svcctx;
memset(ctx->p, 0, sizeof(ctx->p));
LIST_INIT(&ctx->modules);
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
ctx->filters = EB_ROOT;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
appctx->st0 = PROMEX_ST_INIT;
MINOR: applet: Change return value for .init callback function 0 is now returned on success and -1 on error.
2022-05-12 09:52:27 +00:00
return 0;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
/* Callback function that releases a promex applet. This happens when the
* connection with the agent is closed. */
static void promex_appctx_release(struct appctx *appctx)
{
struct promex_ctx *ctx = appctx->svcctx;
struct promex_module_ref *ref, *back;
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
struct promex_metric_filter *flt;
struct eb32_node *node, *next;
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
BUG/MAJOR: promex: fix crash on deleted server Promex applet is used to dump many metrics. Some of them are related to a server instance. The applet can be interrupted in the middle of a dump, for example waiting for output buffer space. In this case, its context is save to resume dump on the correct instance. A crash can occur if dump is interrupted during servers loop. If the server instance is deleted during two scheduling of the promex applet, its context will still referenced the deleted server on resume. To fix this, use server refcount to prevent its deletion during parsing. No backport is needed, despite all stable releases being affected. This is because promex applet context has been recently rewritten to use generic pointers. As such, a specific commit will be applied for earlier releases.
2024-02-22 13:16:37 +00:00
if (appctx->st1 == PROMEX_DUMPER_SRV) {
struct server *srv = objt_server(ctx->p[1]);
srv_drop(srv);
}
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
list_for_each_entry_safe(ref, back, &ctx->modules, list) {
LIST_DELETE(&ref->list);
pool_free(pool_head_promex_mod_ref, ref);
}
MEDIUM: promex: Add support for filters on metric names It is now possible to filter the metrics on their name, by listing explicitly metrics to dump or on the opposite to exclude only some metrics from the dump. To do so, a comma-separated list of metrics must be specified. If a name is preceded by a minus (-), the metric is excluded from the dump. If at least one metric is specified to be explicitly dumped, all metrics are no longer dumped, but only those explicitly listed. The list is specified via one or more "metrics" parameters in the uri query-string. For insance: # Dumped all metrics, except "haproxy_server_check_status" /metrics?metrics=-haproxy_server_check_status # Only dump frontends, backends and servers status /metrics?metrics=haproxy_frontend_status,haproxy_backend_status,haproxy_server_status Included and Excluded metrics can be mixed. Only the intersection will be dumped. This patch should fix the issue #770.
2024-01-31 16:15:49 +00:00
node = eb32_first(&ctx->filters);
while (node) {
next = eb32_next(node);
eb32_delete(node);
flt = container_of(node, typeof(*flt), node);
pool_free(pool_head_promex_metric_flt, flt);
node = next;
}
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
/* The main I/O handler for the promex applet. */
static void promex_appctx_handle_io(struct appctx *appctx)
{
CLEANUP: applet: rename appctx_cs() to appctx_sc() It returns a stream connector, not a conn_stream anymore, so let's fix its name.
2022-05-27 09:08:15 +00:00
struct stconn *sc = appctx_sc(appctx);
CLEANUP: promex: rename all occurrences of stconn "cs" to "sc" Function arguments and local variables called "cs" were renamed to "sc" to avoid future confusion.
2022-05-27 08:16:15 +00:00
struct stream *s = __sc_strm(sc);
struct channel *req = sc_oc(sc);
struct channel *res = sc_ic(sc);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
struct htx *req_htx, *res_htx;
int ret;
res_htx = htx_from_buf(&res->buf);
MEDIUM: promex: Use the sedesc to report and detect end of processing Just like for other applets, we now use the SE descriptor instead of the channel to report error and end-of-stream.
2023-03-31 09:29:54 +00:00
if (unlikely(se_fl_test(appctx->sedesc, (SE_FL_EOS|SE_FL_ERROR|SE_FL_SHR|SE_FL_SHW))))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto out;
CLEANUP: assorted typo fixes in the code and comments This is fifth iteration of typo fixes
2020-03-23 17:28:40 +00:00
/* Check if the input buffer is available. */
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
if (!b_size(&res->buf)) {
MEDIUM: tree-wide: Change sc API to specify required free space to progress sc_need_room() now takes the required free space to receive more data as parameter. All calls to this function are updated accordingly. For now, this value is set but not used. When we are waiting for a buffer, 0 is used. So we expect to be unblocked ASAP. However this must be reviewed because SC_FL_NEED_BUF is probably enough in this case and this flag is already set if the input buffer allocation fails.
2023-05-05 09:28:45 +00:00
sc_need_room(sc, 0);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto out;
}
switch (appctx->st0) {
case PROMEX_ST_INIT:
CLEANUP: promex: rename all occurrences of stconn "cs" to "sc" Function arguments and local variables called "cs" were renamed to "sc" to avoid future confusion.
2022-05-27 08:16:15 +00:00
ret = promex_parse_uri(appctx, sc);
MINOR: contrib/prometheus-exporter: filter exported metrics by scope Now, the prometheus exporter parses the HTTP query-string to filter or to adapt the exported metrics. In this first version, it is only possible select the scopes of metrics to export. To do so, one or more parameters with "scope" as name must be passed in the query-string, with one of those values: global, frontend, backend, server or '*' (means all). A scope parameter with no value means to filter out all scopes (nothing is returned). The scope parameters are parsed in their appearance order in the query-string. So an empty scope will reset all scopes already parsed. But it can be overridden by following scope parameters in the query-string. By default everything is exported. The filtering can also be done on prometheus scraping configuration, but general aim is to optimise the source of data to improve load and scraping time. This is particularly true for huge configuration with thousands of backends and servers. Also note that this configuration was possible on the previous official haproxy exporter but with even more parameters to select the needed metrics. Here we thought it was sufficient to simply avoid a given type of metric. However, more filters are still possible. Thanks to William Dauchy. This patch is based on his work.
2019-11-18 13:47:08 +00:00
if (ret <= 0) {
if (ret == -1)
goto error;
goto out;
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
appctx->st0 = PROMEX_ST_HEAD;
appctx->st1 = PROMEX_DUMPER_INIT;
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
case PROMEX_ST_HEAD:
CLEANUP: promex: rename all occurrences of stconn "cs" to "sc" Function arguments and local variables called "cs" were renamed to "sc" to avoid future confusion.
2022-05-27 08:16:15 +00:00
if (!promex_send_headers(appctx, sc, res_htx))
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
goto out;
appctx->st0 = ((s->txn->meth == HTTP_METH_HEAD) ? PROMEX_ST_DONE : PROMEX_ST_DUMP);
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
case PROMEX_ST_DUMP:
CLEANUP: promex: rename all occurrences of stconn "cs" to "sc" Function arguments and local variables called "cs" were renamed to "sc" to avoid future confusion.
2022-05-27 08:16:15 +00:00
ret = promex_dump_metrics(appctx, sc, res_htx);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
if (ret <= 0) {
if (ret == -1)
goto error;
goto out;
}
appctx->st0 = PROMEX_ST_DONE;
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
case PROMEX_ST_DONE:
BUG/MEDIUM: promex: Be sure to never set EOM flag on an empty HTX message It is the same bug than "BUG/MEDIUM: stats: Be sure to never set EOM flag on an empty HTX message". We must be sure to set the EOM flag on a non empyt message to be sure the mux on the client will properly handle shutdowns. Otherwise, it may miss the end of the message and considers any shutdown as an abort. This patch must be backported as far as 2.4. On previous version there is still the EOM HTX block.
2022-04-07 08:19:46 +00:00
/* no more data are expected. If the response buffer is
* empty, be sure to add something (EOT block in this
* case) to have something to send. It is important to
* be sure the EOM flags will be handled by the
* endpoint.
*/
if (htx_is_empty(res_htx)) {
if (!htx_add_endof(res_htx, HTX_BLK_EOT)) {
MEDIUM: tree-wide: Change sc API to specify required free space to progress sc_need_room() now takes the required free space to receive more data as parameter. All calls to this function are updated accordingly. For now, this value is set but not used. When we are waiting for a buffer, 0 is used. So we expect to be unblocked ASAP. However this must be reviewed because SC_FL_NEED_BUF is probably enough in this case and this flag is already set if the input buffer allocation fails.
2023-05-05 09:28:45 +00:00
sc_need_room(sc, sizeof(struct htx_blk) + 1);
BUG/MEDIUM: promex: Be sure to never set EOM flag on an empty HTX message It is the same bug than "BUG/MEDIUM: stats: Be sure to never set EOM flag on an empty HTX message". We must be sure to set the EOM flag on a non empyt message to be sure the mux on the client will properly handle shutdowns. Otherwise, it may miss the end of the message and considers any shutdown as an abort. This patch must be backported as far as 2.4. On previous version there is still the EOM HTX block.
2022-04-07 08:19:46 +00:00
goto out;
}
channel_add_input(res, 1);
}
res_htx->flags |= HTX_FL_EOM;
CLEANUP: applet: rename the sedesc pointer from "endp" to "sedesc" Now at least it makes it obvious that it's the stream endpoint descriptor and not an endpoint. There were few changes thanks to the previous refactor of the flags.
2022-05-17 16:05:31 +00:00
se_fl_set(appctx->sedesc, SE_FL_EOI);
BUG/MINOR: contrib/prometheus-exporter: Fix applet accordingly to recent changes Since the flag EOI was added on channels, some hidden bugs in the prometheus exporter now leads to error. the visible effect is that responses are truncated. So first of all, channel_add_input() must be called when the response headers and the EOM block are added. To be sure to correctly update the response channel (especially to_forward value). Then the request must really be fully consumed. And finally, the return clause in the switch has been replaced by a break. It was totally wrong to skip the end of the function in the states PROMEX_DONE and PROMEX_ERROR. (Note that PROMEX_ERROR was never used, so it was replaced by PROMEX_END just to ease reading the code). No need to backport this patch, the Prometheus exporter does not exist in early versions.
2019-03-27 14:48:53 +00:00
appctx->st0 = PROMEX_ST_END;
BUILD: prometheus: use __fallthrough in promex_dump_metrics() and IO handler() This avoids 11 build warnings when preprocessing happens before compiling with gcc >= 7.
2022-11-14 06:37:45 +00:00
__fallthrough;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
BUG/MINOR: contrib/prometheus-exporter: Fix applet accordingly to recent changes Since the flag EOI was added on channels, some hidden bugs in the prometheus exporter now leads to error. the visible effect is that responses are truncated. So first of all, channel_add_input() must be called when the response headers and the EOM block are added. To be sure to correctly update the response channel (especially to_forward value). Then the request must really be fully consumed. And finally, the return clause in the switch has been replaced by a break. It was totally wrong to skip the end of the function in the states PROMEX_DONE and PROMEX_ERROR. (Note that PROMEX_ERROR was never used, so it was replaced by PROMEX_END just to ease reading the code). No need to backport this patch, the Prometheus exporter does not exist in early versions.
2019-03-27 14:48:53 +00:00
case PROMEX_ST_END:
MEDIUM: promex: Use the sedesc to report and detect end of processing Just like for other applets, we now use the SE descriptor instead of the channel to report error and end-of-stream.
2023-03-31 09:29:54 +00:00
se_fl_set(appctx->sedesc, SE_FL_EOS);
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
out:
htx_to_buf(res_htx, &res->buf);
BUG/MINOR: contrib/prometheus-exporter: Fix applet accordingly to recent changes Since the flag EOI was added on channels, some hidden bugs in the prometheus exporter now leads to error. the visible effect is that responses are truncated. So first of all, channel_add_input() must be called when the response headers and the EOM block are added. To be sure to correctly update the response channel (especially to_forward value). Then the request must really be fully consumed. And finally, the return clause in the switch has been replaced by a break. It was totally wrong to skip the end of the function in the states PROMEX_DONE and PROMEX_ERROR. (Note that PROMEX_ERROR was never used, so it was replaced by PROMEX_END just to ease reading the code). No need to backport this patch, the Prometheus exporter does not exist in early versions.
2019-03-27 14:48:53 +00:00
/* eat the whole request */
if (co_data(req)) {
req_htx = htx_from_buf(&req->buf);
co_htx_skip(req, req_htx, co_data(req));
}
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
return;
error:
MEDIUM: promex: Use the sedesc to report and detect end of processing Just like for other applets, we now use the SE descriptor instead of the channel to report error and end-of-stream.
2023-03-31 09:29:54 +00:00
se_fl_set(appctx->sedesc, SE_FL_ERROR);
BUG/MINOR: promex: Don't forget to consume the request on error When the promex applet triggers an error, for instance because the URI is invalid, we must still take care to consume the request. Otherwise, the error will be handled by HTTP analyzers as a server abort. This patch must be backported as far as 2.0.
2023-01-13 07:53:23 +00:00
goto out;
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
}
struct applet promex_applet = {
.obj_type = OBJ_TYPE_APPLET,
.name = "<PROMEX>", /* used for logging */
.init = promex_appctx_init,
MEDIUM: promex: Dump metrics of registered modules with a way to filter them This patch adds a dump loop on the registered modules. It is very similar to other dump loops. When a module registered, a implicit scope is created with the module's name. It means a module name must be unique. It also means, metrics dump of modules can be filtered via the "scope" parameter.
2024-01-30 14:58:31 +00:00
.release = promex_appctx_release,
CONTRIB: contrib/prometheus-exporter: Add a Prometheus exporter for HAProxy It has been developped as a service applet. Internally, it is called "promex". To build HAProxy with the promex service, you should use the Makefile variable "EXTRA_OBJS". To be used, it must be enabled in the configuration with an "http-request" rule and the corresponding HTTP proxy must enable the HTX support. For instance: frontend test mode http ... option http-use-htx http-request use-service prometheus-exporter if { path /metrics } ... See contrib/prometheus-exporter/README for details.
2019-02-07 14:38:42 +00:00
.fct = promex_appctx_handle_io,
};
static enum act_parse_ret service_parse_prometheus_exporter(const char **args, int *cur_arg, struct proxy *px,
struct act_rule *rule, char **err)
{
/* Prometheus exporter service is only available on "http-request" rulesets */
if (rule->from != ACT_F_HTTP_REQ) {
memprintf(err, "Prometheus exporter service only available on 'http-request' rulesets");
return ACT_RET_PRS_ERR;
}
/* Add applet pointer in the rule. */
rule->applet = promex_applet;
return ACT_RET_PRS_OK;
}
static void promex_register_build_options(void)
{
char *ptr = NULL;
memprintf(&ptr, "Built with the Prometheus exporter as a service");
hap_register_build_opts(ptr, 1);
}
static struct action_kw_list service_actions = { ILH, {
{ "prometheus-exporter", service_parse_prometheus_exporter },
{ /* END */ }
}};
INITCALL1(STG_REGISTER, service_keywords_register, &service_actions);
INITCALL0(STG_REGISTER, promex_register_build_options);
Reference in New Issue Copy Permalink