hub: optional startupProbe for slow-startup migration windows

charts/lunar/CHANGELOG.md

@@ -9,6 +9,19 @@ History starts at 1.0.0 (the snippet→script rename and ghcr.io
 switchover); earlier 0.x versions had no production users. For 0.x
 history see `git log -- charts/lunar/`.
 
+## [Unreleased]
+
+### Added
+
+- **Optional `hub.startupProbe`.** New optional probe (disabled by default
+  to preserve current behaviour) that gives Hub a longer window to come
+  Ready before liveness/readiness kick in. Useful when Hub takes longer
+  than `livenessProbe.failureThreshold × periodSeconds` to start — e.g.
+  on a cold start with a substantial Postgres schema migration.
+  Default tuning when enabled: `30 × 5s = 150s` startup window. Set
+  `hub.startupProbe.enabled: true` and tune `failureThreshold` /
+  `periodSeconds` to taste.
+
 ## [2.1.0] - 2026-05-25
 
 ### Added

charts/lunar/templates/hub-deployment.yaml

@@ -201,6 +201,17 @@ spec:
             failureThreshold: {{ .failureThreshold }}
           {{- end }}
           {{- end }}
+          {{- if .Values.hub.startupProbe.enabled }}
+          {{- with .Values.hub.startupProbe }}
+          startupProbe:
+            httpGet:
+              path: /health
+              port: hub-health
+            initialDelaySeconds: {{ .initialDelaySeconds }}
+            periodSeconds: {{ .periodSeconds }}
+            failureThreshold: {{ .failureThreshold }}
+          {{- end }}
+          {{- end }}
           ports:
             - name: hub-grpc
               containerPort: 8000

charts/lunar/values.yaml

@@ -279,6 +279,22 @@ hub:
     periodSeconds: 5
     failureThreshold: 3
 
+  # Optional startup probe (disabled by default to preserve existing
+  # behaviour). Enable when Hub takes longer to become Ready than the
+  # liveness probe's failureThreshold × periodSeconds tolerates — e.g.
+  # on a cold start with a substantial Postgres schema migration. While
+  # the startupProbe is running, kubelet suppresses liveness/readiness
+  # probe failures, so Hub gets `failureThreshold × periodSeconds`
+  # seconds total before kubelet starts considering it unhealthy.
+  #
+  # Default tuning when enabled is 30 × 5s = 150s. Raise
+  # `failureThreshold` for larger databases with longer migrations.
+  startupProbe:
+    enabled: false
+    initialDelaySeconds: 0
+    periodSeconds: 5
+    failureThreshold: 30
+
   labels: {}
   annotations: {}
   podAnnotations: {}