Readiness probe fails due to missing /api/public/ready endpoint
warningavailabilityUpdated Mar 7, 2026(via Exa)
Technologies:
How to detect:
When deploying Langfuse with Helm chart version 1.5.22 and app version 3.155.1, the readiness probe configured to check /api/public/ready returns HTTP 404, causing web pods to remain in 0/1 Ready state indefinitely despite the application being functional. Application logs show 'Ready in 959ms' but Kubernetes never routes traffic due to failed readiness checks.
Recommended action:
Update values.yaml to change readinessProbe.path from '/api/public/ready' to '/api/public/health' (same endpoint used by liveness probe). Apply changes with 'helm upgrade langfuse langfuse/langfuse --namespace langfuse --values langfuse-helm-values.yaml'. Verify pods reach 1/1 Ready state with 'kubectl get pods -n langfuse'.