Project

General

Profile

Actions
Copy link

Documentation #52656

closed

mgr/prometheus: wron unit in RBD latency metric description

Added by Konstantin Shalygin over 2 years ago. Updated 8 months ago.

Status:
Resolved
Priority:
Low
Assignee:
Konstantin Shalygin
Category:
prometheus module
Target version:
Ceph - v19.0.0
% Done:

100%

Tags:
Backport:
Reviewed:
Affected Versions:
Ceph - v14.2.22, Ceph - v15.2.14, Ceph - v16.2.6
Pull request ID:
43217

Description

For _write_latency and _read_latency metrics the unit in description is msec, and users who makes dashboards was discouraged when see latencies for disks in hours, because actual unit should be nsec

Screenshot from @ceph_ru/@cinder2 user

Files

screenshot_from_2021-09-14_16-25-10.png (70.2 KB) screenshot_from_2021-09-14_16-25-10.png Screenshot from @ceph_ru/@cinder2 user Konstantin Shalygin, 09/18/2021 10:39 AM
Actions
Copy link
 #1

Updated by Konstantin Shalygin over 2 years ago

  • Tracker changed from Bug to Documentation
  • Status changed from New to In Progress
  • Pull request ID set to 43217
Actions
Copy link
 #2

Updated by Anthony D'Atri over 2 years ago

I've looked at `monitoring/grafana/dashboards/rbd-overview.json` and `monitoring/grafana/dashboards/rbd-overview.json` and in both the Y axis unit appears to be set to `ns` in both.

But since the pane titles here are "Average Latency", I suspect that those aren't what you're talking about.

`src/pybind/mgr/prometheus/module.py` however does describe these metrics in terms of ms so I suspect that's what you ran into.
https://github.com/ceph/ceph/pull/43479

Actions
Copy link
 #3

Updated by Konstantin Shalygin 8 months ago

  • Status changed from In Progress to Resolved
  • Target version changed from v17.0.0 to v19.0.0
  • % Done changed from 0 to 100
  • Backport deleted (pacific octopus)
Actions
Copy link

Also available in: Atom PDF