RepoMirrors
/
ceph_exporter
mirror of https://github.com/digitalocean/ceph_exporter
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
175 Commits 10 Branches 24 Tags 1.3 MiB
Commit Graph

20 Commits

Author SHA1 Message Date
Sasha Krutiy 1a3dff593e health: added health check warnings map into collector, removed from config 2019-09-09 11:45:17 -04:00
Sasha Krutiy 8b3221a76f exporter: fixed type errors with struct for storing health warnings map 2019-09-06 15:17:39 -04:00
Sasha Krutiy 5e40382bad Added list of ceph health checks and their criticality, and the setting of status based on these checks 2019-09-05 11:57:21 -04:00
Vaibhav Bhembre 4cd78e49f0 exporter: register pool info collector 2019-05-17 14:03:29 -04:00
ssobolewski 415d296c31
Ssobolewski/run rgw stats in background (#97) 
* RGW GC stat collection can take a long time if there is a very large backlog

* Use a const for background interval

* Minor change per code review
 2018-08-10 13:43:02 -06:00
ssobolewski dc6ab9c636
Optionally collect RGW GC task stats (#94) 
* Optionally collect RGW GC task stats

* Minor changes per code-review, add some additional tests to squeeze out extra coverage
 2018-08-01 07:37:07 -06:00
Jan Fajerski a42a258a28 Add constant for tcp keepalive periode 
Signed-off-by: Jan Fajerski <jfajerski@suse.com>
 2018-07-09 11:08:14 +02:00
Tim Serong cd9aa031a8 Terminate exporter process if maximum open files exceeded 
This is somewhat of a workaround for the exporter becoming
perpetually blocked when it runs out of file descriptors if
the cluster is down for too long, as mentioned in:

  https://github.com/digitalocean/ceph_exporter/issues/60#issuecomment-319396108

The problem is that if the MONs are down for long enough,
each time prometheus scrapes the metrics, another socket is
opened, but these block forever.  If the cluster comes back
up before we run out of FDs, the blocked requests recover.
If the clusetr *doesn't* come back up before we run out of
FDs, the blocked requests never recover.

This commit causes ceph exporter to terminate if it runs
out of file descriptors, which IMO is better than blocking
forever -- it'll be a noisier failure, and also if you're
running ceph_exporter via systemd, systemd will then
automatically trigger a service restart.

Signed-off-by: Tim Serong <tserong@suse.com>
(cherry picked from commit bb1ad364b5)
 2018-06-12 10:49:04 +02:00
utkarshmani1997 4c70969940 update /metrics handler to promhttp.Handler() 2017-09-14 17:09:52 +05:30
ssobolewski f67a9a0f7f Add multicluster support to ceph_exporter. This change allows one instance to monitor multipl ceph clusters and report to the same prometheus instance 2017-03-28 15:58:27 -06:00
Vaibhav Bhembre 2d86f232a4 travis: fix docker push, add default page 2016-12-29 18:00:50 -05:00
root e921602291 Remove unecessary flag usage 2016-12-28 14:31:37 +00:00
root 5451d4c28d Gofmt 2016-12-28 14:27:54 +00:00
root 3875fb97bb CSupport ceph user other than admin 2016-12-28 12:56:41 +00:00
madzfe a548fe25b4 Osd to OSD. 2016-07-29 11:04:55 +08:00
madzfe 335e7be325 change import back to digitalocean 2016-07-26 11:19:26 +08:00
madz 4699d24605 add osd metrics in ceph exporter 2016-07-14 13:01:19 +08:00
Vaibhav Bhembre 7efbf4c9eb Pick a default prometheus port for CephExporter 2016-01-29 14:45:17 +00:00
Matt Layher 526d3fb634 Add doc comments, clarify error log messages, use collectors package name 2016-01-10 15:36:54 -05:00
Vaibhav Bhembre 9cda67d44a Add Ceph Exporter 2016-01-06 13:27:16 -05:00