fix: protect Authentik server startup after database recovery #1
No reviewers
Labels
No labels
bug
documentation
duplicate
enhancement
good first issue
help wanted
invalid
question
wontfix
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
homelab/authentik-operator!1
Loading…
Reference in a new issue
No description provided.
Delete branch "fix/3f88f077-startup-probe"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Summary
Give Authentik server ten minutes to start after PostgreSQL and its migration worker recover from a reboot. Kubernetes gates the existing liveness/readiness probes on an HTTP startup probe:
/-/health/live/, port 9000, 60 × 10 seconds, five-second timeout.The main release workflow publishes the explicitly reviewed chart version 0.1.2 and refuses an existing tag. This avoids deriving the already published 0.1.0 from stale v0.0.15 tag history. Release/fixture documentation changes in the same commits as their behavior.
Part of HOMEIT-368. Deploy postgres ownership repair first; base documentation/pointer follows this PR.
Evidence
a8c93fc: unit tests, lint, Kind E2E. All four E2E specs pass; the run builds/starts the operator image, verifies the generated startup budget and reconciles a real server/client.Merge danger
Door: one-way publication. The mirror's main workflow automatically publishes v0.1.2, multiarchitecture images, a release, install manifest and Helm chart. Morten's release approval is required before merge. Publication does not change core's live pin.
Blast radius: household. Deploying the new operator reconciles Authentik and rolls its server, which provides login to NetBird and Coder. The workload change is reversible.
Rollback and verification