Page MenuHomeDevCentral

Restore metrics collector in management plugin
ClosedPublic

Authored by dereckson on May 30 2023, 19:24.
Tags
None
Referenced Files
Unknown Object (File)
Fri, Nov 22, 12:17
Unknown Object (File)
Fri, Nov 22, 12:06
Unknown Object (File)
Fri, Nov 22, 05:59
Unknown Object (File)
Sun, Nov 17, 18:48
Unknown Object (File)
Fri, Nov 15, 17:59
Unknown Object (File)
Fri, Nov 15, 17:58
Unknown Object (File)
Fri, Nov 15, 17:58
Unknown Object (File)
Fri, Nov 15, 17:38
Subscribers
None

Details

Summary

The base image disable explicitly fot the management plugin
the metrics collector and associated UI to avoid redundancy
with Prometheus metrics.

Metrics in the UI for queues is something useful to debug
in an unified place, even with dashboards on the side.

For a discussion on metrics collection, see also T1623 and T1633.

Test Plan

Deploy for white-rabbit and orange-rabbit

Diff Detail

Repository
rDRABBITMQ RabbitMQ Docker image
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

dereckson created this revision.

Still "Stats in management UI are disabled on this node"

This revision is now accepted and ready to land.May 30 2023, 19:27
dereckson added inline comments.
Dockerfile
7

with BuildKit I'm not sure that's a good idea to remove a file

Hmmm they prioritized the configuration to load in a specific order, so now it's 20-management_agent.disable_metrics_collector.conf

They change again the config, BOOM we lose it again.

Regardless of the previous remark on the lack of stability of upstream image, we can currently proceed.

Update configuration file name

This revision is now accepted and ready to land.May 30 2023, 19:34