Project

General

Profile

Actions
Copy link

Bug #61572

open

mgr: remove invalid zero performance counter

Added by Zhansong Gao 11 months ago. Updated 11 months ago.

Status:
Fix Under Review
Priority:
Normal
Assignee:
Zhansong Gao
Category:
prometheus module
Target version:
Ceph - v19.0.0
% Done:

0%

Source:
Community (dev)
Tags:
Backport:
reef quincy pacific
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Pull request ID:
51883
Crash signature (v1):
Crash signature (v2):

Description

The Prometheus server may get the following data when a counter has no data available for a period of time:

'10100 @1685548800.000'
'10200 @1685548815.000'
'0 @1685548830.000'
'10400 @1685548845.000'

Assume this counter is 'ceph_osd_op_r' and we will get an incorrect value when calculating 'rate(ceph_osd_op_r[1m])'.

Actions
Copy link
 #1

Updated by Konstantin Shalygin 11 months ago

  • Category set to prometheus module
  • Status changed from New to Fix Under Review
  • Assignee set to Zhansong Gao
  • Target version set to v19.0.0
  • Source set to Community (dev)
  • Backport set to reef quincy pacific
Actions
Copy link

Also available in: Atom PDF