Project

General

Profile

Feature #56179

[RFE] Our prometheus instance should scrape itself

Added by Paul Cuzner 6 months ago. Updated 6 months ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
cephadm/monitoring
Target version:
% Done:

0%

Source:
Tags:
Backport:
Reviewed:
Affected Versions:
Pull request ID:

Description

At the moment we don't scrape metrics from the prometheus server itself.

If we did, we could
  • track and monitor prometheus growth (rate of expansion, and maybe use predict linear to forecast)
  • observe scrape sample sizes and growth over time - look for anomalies
  • alert based on issues in the tsdb/wal
  • pick up and alert on cephadm not responding to sd refresh requests
    etc etc

History

#1 Updated by Arthur Outhenin-Chalandre 6 months ago

I don't think it covers everything you want but Prometheus has a mixin: https://github.com/prometheus/prometheus/tree/main/documentation/prometheus-mixin but I guess you could also try to add the missing piece there as well.

#2 Updated by Redouane Kachach Elhichou 6 months ago

Thanks for writing down the RFE proposal. Please, can you put some more information about what config needs to be add in order to support this new feature. Any pointers to the docs are also more than welcome.

Also available in: Atom PDF