Prometheus: replace default scrapeConfigs to fix duplicate scrape job

Lucas Lirequested to merge
lli/prometheus-fix-duplicate-scrape-job into master

What does this MR do?

The bundled Prometheus subchart (29.17.0) renders scrape configs from two additive sources: its own populated scrapeConfigs map (~10 default jobs) and GitLab's legacy serverFiles."prometheus.yml".scrape_configs. GitLab attempted to suppress the upstream defaults with scrapeConfigs: null, but null is silently dropped when coalescing into the condition-gated (prometheus.install) subchart (helm/helm#31446, helm/helm#9136), so the upstream defaults survived and collided with GitLab's own jobs, causing Prometheus to fail to load with found multiple scrape configs with job name "prometheus".

Replace the ineffective null with an empty list, scrapeConfigs: []. A non-null value propagates through the condition-gated coalesce, and a list-typed value replaces the upstream default map wholesale (list-vs-map type mismatch), leaving only GitLab's own serverFiles jobs. Rendering now emits exactly the 5 GitLab jobs, each once, with no duplicates, no placeholder job, and no Helm coalesce warning.

Interim fix; it relies on Helm's list-replaces-map coalesce behavior. The durable fix (migrating GitLab's jobs onto the upstream scrapeConfigs structure) is tracked in #6413.

Related issues

#6413

Author checklist

For general guidance, please follow our Contributing guide.

Required

For anything in this list which will not be completed, please provide a reason in the MR discussion.

  • Merge Request Title and Description are up to date, accurate, and descriptive.
  • MR targeting the appropriate branch.
  • MR has a green pipeline.
  • Documentation created/updated.
  • Tests added/updated.
  • Have you validated that your change works end-to-end on a real cluster deployment?
    • Describe how you did it in a comment.
  • Equivalent MR/issue for omnibus-gitlab opened.

Reviewers checklist

Edited by Clemens Beck

Merge request reports