Project

General

Profile

Bug #49443

mgr/prometheus doesn't expose used_bytes per pool

Added by Paul Cuzner about 2 months ago. Updated about 1 month ago.

Status:
Closed
Priority:
Normal
Assignee:
Category:
prometheus module
Target version:
% Done:

0%

Source:
Tags:
Backport:
pacific octopus nautilus
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
Crash signature (v1):
Crash signature (v2):

Description

The current metrics exposed don't show actual space consumed within a pool (used_bytes) - so you have less data available to prometheus for alerting/monitoring than you do at the command line with ceph df detail

History

#1 Updated by Konstantin Shalygin about 2 months ago

Can you explain?
Currently exporter have `ceph_pool_bytes_used` metric.

#2 Updated by Paul Cuzner about 1 month ago

  • Status changed from New to Closed

Konstantin Shalygin wrote:

Can you explain?
Currently exporter have `ceph_pool_bytes_used` metric.

Apologies for the delay.
the parms in DF_POOL omit the used_bytes field, so although we can report in stored (ingested data), we can't see the actual consumed (used_bytes), which includes the effect of compression/ec/replication overheads

I'll be closing this tracker, and include it with a related feature tracker #49049 (https://tracker.ceph.com/issues/49049)

Also available in: Atom PDF