Bug #64295: Ceph exporter does not produce usable RGW metrics in k8 envs - Ceph - Ceph

Actions

Copy link

Bug #64295

open

Ceph exporter does not produce usable RGW metrics in k8 envs

Added by Juan Miguel Olmo Martínez 3 months ago. Updated 3 months ago.

Status:

New

Priority:

Normal

Assignee:

Avan Thakkar

Category:

Monitoring/Alerting

Target version:

v18.2.3

% Done:

Source:

Tags:

Backport:

Regression:

Severity:

2 - major

Reviewed:

Affected Versions:

ceph-qa-suite:

Pull request ID:

Crash signature (v1):

Crash signature (v2):

Description

The way Ceph exporter links RGW metrics coming from RGW daemons with the RGW daemon metadata metrics coming from Prometheus manager module is not working properly in k8s envs (rook).
Ceph exporter extracts the identification of the daemon from the daemon admin socket file name. This file name uses different patterns depending if the daemon is deployed in a baremetal env(cephadm clusters) or k8s env (rook clusters), the current implementation only covers cephadm clusters.

This cause that RGW metrics will not be usable in k8s envs

Note about current implementation:
https://github.com/ceph/ceph/blob/3cb0412e26db5b5a7c92be9e9cbeff3813345caa/src/exporter/DaemonMetricCollector.cc#L291
Probably using regular expressions instead fixed string positions will provide more robustness to the code.

Related issues 1 (1 open — 0 closed)