Skip to content

Implement liveness check#331

Draft
victor-girardin-unity wants to merge 1 commit intoredpanda-data:masterfrom
victor-girardin-unity:master
Draft

Implement liveness check#331
victor-girardin-unity wants to merge 1 commit intoredpanda-data:masterfrom
victor-girardin-unity:master

Conversation

@victor-girardin-unity
Copy link
Copy Markdown

Previously there was no liveness probe on the Deployment, and the DaemonSet reused /ready for both probes. In OffsetsTopic scrape mode, /ready returns 503 until the consumer catches up on __consumer_offsets, which can take minutes on large clusters. Using /ready for liveness would cause Kubernetes to kill pods that are still starting up. The new /live endpoint avoids this by only checking that the HTTP server is responsive.

Work

  • Add /live HTTP endpoint that always returns 200, separating liveness from readiness concerns
  • Add liveness probe configuration to the Helm Deployment template (enabled by default)
  • Update DaemonSet template to use /live instead of reusing /ready for liveness

@victor-girardin-unity victor-girardin-unity marked this pull request as draft March 18, 2026 15:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant