fix(homarr): add proper probes via Kustomize patch
Chart version 8.23.0 doesn't support probe customization via values. Using strategic merge patch instead. Probe configuration: - Liveness: 60s initial delay, 30s period, 5s timeout - Readiness: 45s initial delay, 15s period, 5s timeout App initialization timeline: 0-5s: DB migrations, Redis startup 6s: WebSocket server ready 27-30s: Icon cache populated (27k+ icons) 30s+: Analytics cron initialized, fully operational
This commit is contained in:
@@ -8,30 +8,13 @@ image:
|
||||
|
||||
replicaCount: 1
|
||||
|
||||
# Override default probes - chart defaults are too aggressive for startup time
|
||||
probes:
|
||||
liveness:
|
||||
enabled: true
|
||||
custom: true
|
||||
spec:
|
||||
httpGet:
|
||||
path: /api/health/live
|
||||
port: 7575
|
||||
initialDelaySeconds: 60 # App needs time for migrations + icon fetch
|
||||
periodSeconds: 30
|
||||
timeoutSeconds: 5
|
||||
failureThreshold: 3
|
||||
readiness:
|
||||
enabled: true
|
||||
custom: true
|
||||
spec:
|
||||
httpGet:
|
||||
path: /api/health/ready
|
||||
port: 7575
|
||||
initialDelaySeconds: 45 # Wait for WebSocket + analytics cron
|
||||
periodSeconds: 15
|
||||
timeoutSeconds: 5
|
||||
failureThreshold: 3
|
||||
# Disable default probes - they kill the container before it finishes startup
|
||||
# Will add custom probes via Kustomize patch with proper timing
|
||||
livenessProbe:
|
||||
enabled: false
|
||||
|
||||
readinessProbe:
|
||||
enabled: false
|
||||
|
||||
tolerations:
|
||||
- key: node-role.kubernetes.io/control-plane
|
||||
|
||||
Reference in New Issue
Block a user