Prometheus metrics are missing for sidekiq processes
Prometheus metrics don't seem to be getting scraped consistently for all sidekiq processes in a sidekiq cluster.
For example, this change is no longer working as expected: #32258 (closed)
Charting this metric over 1 day:
sidekiq_concurrency{fqdn="sidekiq-pipeline-01-sv-gprd.c.gitlab-production.internal"}
The problem appears to be that the results for a single pid
are being returned, not all pid
s in the sidekiq cluster on the host.
cc @mkaeppler
Edited by 🤖 GitLab Bot 🤖