fix(api,task-processor): Close shutdown drain gap and name container ports

charts/flagsmith/templates/deployment-api.yaml

@@ -126,7 +126,12 @@ spec:
         {{- end }}
         {{- end }}
         ports:
-        - containerPort: {{ .Values.service.api.port }}
+        - name: http
+          containerPort: {{ .Values.service.api.port }}
+        {{- if .Values.prometheus.enabled }}
+        - name: prom
+          containerPort: 9100
+        {{- end }}
         env: {{ include (print $.Template.BasePath "/_api_environment.yaml") . | nindent 8 }}
         livenessProbe:
           {{- if .Values.api.livenessProbe.exec }}

charts/flagsmith/templates/deployment-task-processor.yaml

@@ -93,7 +93,12 @@ spec:
           - run-task-processor
         {{- end }}
         ports:
-        - containerPort: {{ .Values.service.taskProcessor.port }}
+        - name: http
+          containerPort: {{ .Values.service.taskProcessor.port }}
+        {{- if .Values.prometheus.enabled }}
+        - name: prom
+          containerPort: 9100
+        {{- end }}
         env: {{ include (print $.Template.BasePath "/_task_processor_environment.yaml") . | nindent 8 }}
         livenessProbe:
           {{- $exec := .Values.taskProcessor.livenessProbe.exec | default .Values.api.livenessProbe.exec }}

charts/flagsmith/values.yaml

@@ -74,10 +74,19 @@ api:
   nodeSelector: {}
   tolerations: []
   affinity: {}
-  # Container lifecycle hooks (e.g. preStop for graceful shutdown)
-  lifecycle: {}
-  # Pod termination grace period in seconds
-  terminationGracePeriodSeconds: null
+  # Container lifecycle hooks. Default preStop delays SIGTERM so the
+  # LB / endpoints controller has time to deregister the pod before
+  # gunicorn closes its listen socket. Without this, rolling deploys
+  # and HPA scale-down can cause a short 5xx spike on traffic that
+  # the LB routes to the pod after it has stopped accepting connections.
+  lifecycle:
+    preStop:
+      exec:
+        command: ["sleep", "20"]
+  # Pod termination grace period in seconds. Must exceed the LB's
+  # connection-draining timeout so the kubelet does not SIGKILL
+  # the pod while the LB is still draining in-flight connections.
+  terminationGracePeriodSeconds: 75
   podSecurityContext: {}
   defaultPodSecurityContext:
     enabled: true
@@ -461,7 +470,7 @@ hpa:
     enabled: false
     minReplicas: 2
     maxReplicas: 10
-    targetCPUUtilization: 50 
+    targetCPUUtilization: 50
   sse:
     enabled: true
     minReplicas: 1