fix(shield): add startupProbe to cluster-shield deployment

charts/shield/Chart.yaml

@@ -13,5 +13,5 @@ maintainers:
   - name: mavimo
     email: [email protected]
 type: application
-version: 1.34.5
+version: 1.34.6
 appVersion: "1.0.0"

charts/shield/templates/cluster/deployment.yaml

@@ -132,6 +132,13 @@ spec:
             {{- include "cluster.env" . | nindent 12 }}
           resources:
             {{- toYaml .Values.cluster.resources | nindent 12 }}
+          {{- if .Values.cluster.probes.startup }}
+          startupProbe:
+            httpGet:
+              path: /healthz
+              port: {{ .Values.cluster.additional_settings.monitoring_port }}
+            {{- .Values.cluster.probes.startup | toYaml | nindent 12 }}
+          {{- end }}
           livenessProbe:
             httpGet:
               path: /healthz

charts/shield/tests/cluster/deployment_test.yaml

@@ -256,6 +256,52 @@ tests:
             periodSeconds: 5
     template: templates/cluster/deployment.yaml
 
+  - it: Default startup probe
+    asserts:
+      - equal:
+          path: spec.template.spec.containers[?(@.name == "cluster-shield")].startupProbe
+          value:
+            httpGet:
+              path: /healthz
+              port: 8080
+            failureThreshold: 30
+            initialDelaySeconds: 15
+            periodSeconds: 10
+            timeoutSeconds: 3
+    template: templates/cluster/deployment.yaml
+
+  - it: Custom startup probe
+    set:
+      cluster:
+        probes:
+          startup:
+            failureThreshold: 5
+            initialDelaySeconds: 2
+            periodSeconds: 3
+            timeoutSeconds: 1
+    asserts:
+      - equal:
+          path: spec.template.spec.containers[?(@.name == "cluster-shield")].startupProbe
+          value:
+            httpGet:
+              path: /healthz
+              port: 8080
+            failureThreshold: 5
+            initialDelaySeconds: 2
+            periodSeconds: 3
+            timeoutSeconds: 1
+    template: templates/cluster/deployment.yaml
+
+  - it: Disabled startup probe
+    set:
+      cluster:
+        probes:
+          startup: null
+    asserts:
+      - notExists:
+          path: spec.template.spec.containers[?(@.name == "cluster-shield")].startupProbe
+    template: templates/cluster/deployment.yaml
+
   - it: Custom readiness probe
     set:
       cluster:

charts/shield/values.yaml

@@ -537,6 +537,18 @@ cluster:
       periodSeconds: 5
       # The readiness probe failure threshold
       failureThreshold: 9
+    startup:
+      # The startup probe initial delay
+      initialDelaySeconds: 15
+      # The startup probe period
+      periodSeconds: 10
+      # The startup probe failure threshold. Combined with periodSeconds, it gives
+      # the container ~5 minutes to finish warmup before liveness takes over. This
+      # protects slow-booting clusters (e.g. GKE Windows) from being killed by the
+      # liveness probe during initial health-check warmup.
+      failureThreshold: 30
+      # The startup probe timeout
+      timeoutSeconds: 3
   # The number of replicas for the cluster shield
   replica_count: 2
   update_strategy: