improvement: add podmonitor for vpc metric collection

charts/aws-vpc-cni/templates/daemonset.yaml

@@ -128,6 +128,9 @@ spec:
         - name: aws-eks-nodeagent
           image: {{ include "aws-vpc-cni.nodeAgentImage" . }}
           imagePullPolicy: {{ .Values.nodeAgent.image.pullPolicy }}
+          ports:
+            - containerPort: {{ .Values.nodeAgent.metricsBindAddr}}
+              name: agentmetrics
           env:
             - name: MY_NODE_NAME
               valueFrom:

charts/aws-vpc-cni/templates/podmonitor.yaml

@@ -0,0 +1,40 @@
+{{- if .Values.podMonitor.create }}
+apiVersion: monitoring.coreos.com/v1
+kind: PodMonitor
+metadata:
+  name: {{ include "aws-vpc-cni.fullname" . }}
+  namespace: {{ .Release.Namespace }}
+  labels:
+    {{- with .Values.podMonitor.labels }}
+    {{- toYaml . | nindent 4 }}
+    {{- end }}
+  {{- with .Values.podMonitor.annotations }}
+  annotations:
+    {{- toYaml . | nindent 4 }}
+  {{- end }}
+spec:
+  jobLabel: {{ include "aws-vpc-cni.fullname" . }}
+  namespaceSelector:
+    matchNames:
+      - {{ .Release.Namespace }}
+  podMetricsEndpoints:
+  - interval: {{ .Values.podMonitor.interval }}
+    path: /metrics
+    port: metrics
+    {{- with .Values.podMonitor.relabelings }}
+    relabelings:
+      {{- toYaml . | nindent 6 }}
+    {{- end }}
+  {{- if .Values.nodeAgent.enabled }}
+  - interval: {{ .Values.podMonitor.interval }}
+    path: /metrics
+    port: agentmetrics
+    {{- with .Values.podMonitor.relabelings }}
+    relabelings:
+      {{- toYaml . | nindent 6 }}
+    {{- end }}
+  {{- end }}
+  selector:
+    matchLabels:
+      k8s-app: aws-node
+{{- end }}

charts/aws-vpc-cni/values.yaml

@@ -46,6 +46,7 @@ nodeAgent:
   networkPolicyAgentLogFileLocation: "/var/log/aws-routed-eni/network-policy-agent.log"
   enableIpv6: "false"
   metricsBindAddr: "8162"
+  metricsBindPort: "8162"
   healthProbeBindAddr: "8163"
   conntrackCacheCleanupPeriod: 300
   resources: {}
@@ -229,3 +230,17 @@ eniConfig:
     #   id: subnet-789
     #   securityGroups:
     #   - sg-789
+
+podMonitor:
+  # Create Prometheus podMonitor 
+  create: false
+  # Annotations to add to the Prometheus podMonitor
+  annotations: {}
+  # Labels to add to the Prometheus podMonitor
+  labels: {}
+  # The interval to scrape metrics.
+  interval: 30s
+  # The timeout before a metrics scrape fails.
+  scrapeTimeout: 30s
+  # relabelings to apply to the podMonitor
+  relabelings: []