Enabling Observability for cilium-agent in a Cluster with DataPlane V2 Enabled

Prerequisites

• CCE DataPlane V2 is released with restrictions. To use this feature, submit a service ticket to CCE.
• Currently, only CCE standard clusters of v1.27.16-r50, v1.28.15-r40, v1.29.15-r0, v1.30.14-r0, v1.31.10-r0, v1.32.6-r0, and later versions that have DataPlane V2 enabled support the observability of the cilium-agent component.

Step 1: Configure a ConfigMap and Rebuild yangtse-cilium

1. Connect to your cluster using kubectl. For details, see Accessing a Cluster Using kubectl.
2. Create the cilium-config.yaml file and specify the following native ConfigMap configuration of the Cilium community:

   apiVersion: v1
   data:
     prometheus-serve-addr: :9962
   kind: ConfigMap
   metadata:
     name: cilium-config
     namespace: kube-system

   Table 1 Parameters

   Parameter	Description	Remarks
   prometheus-serve-addr	Address for accessing the metrics of the cilium-agent component.	This parameter is set to :9962 here. The colon cannot be omitted.

3. Create the ConfigMap.

   kubectl apply -f cilium-config.yaml

4. Roll back and rebuild yangtse-cilium to apply the configuration.

   uuid=$(uuidgen)
   kubectl patch daemonset -nkube-system yangtse-cilium --type='json' -p="[{\"op\": \"add\", \"path\": \"/spec/template/metadata/annotations/change-id\", \"value\": \"$uuid\"}]"

Step 2: Install Add-ons

Step 3: Configure a Cilium Observability Task

1. Create the cilium-monitor.yaml file. You can specify the following configuration to create a ServiceMonitor for Cilium observability.

   apiVersion: v1
   kind: Service
   metadata:
     annotations:
       prometheus.io/port: "9962"
       prometheus.io/scrape: "true"
     labels:
       k8s-app: cilium
     name: cilium
     namespace: kube-system
   spec:
     clusterIP: None
     clusterIPs:
     - None
     internalTrafficPolicy: Cluster
     ipFamilies:
     - IPv4
     ipFamilyPolicy: SingleStack
     ports:
     - name: metrics
       port: 9962
       protocol: TCP
       targetPort: prometheus
     selector:
       k8s-app: cilium
     sessionAffinity: None
     type: ClusterIP
   ---
   apiVersion: monitoring.coreos.com/v1
   kind: ServiceMonitor
   metadata:
     name: cilium
     namespace: monitoring
   spec:
     endpoints:
     - honorLabels: true
       path: /metrics
       port: metrics
       relabelings:
       - action: labelmap
         regex: __meta_kubernetes_service_label_(.+)
       scheme: http
     jobLabel: cilium
     namespaceSelector:
       matchNames:
       - kube-system
     selector:
       matchLabels:
         k8s-app: cilium

2. Create the ServiceMonitor.

   kubectl apply -f cilium-monitor.yaml

Step 4: Configure the Cilium Metric Dashboard

The Cilium official website provides a Grafana dashboard for Cilium metrics. Compared with the monotonous metrics of Prometheus, the Grafana dashboard is more attractive. You can access the dashboard to view Cilium metrics.

1. Prepare the configuration file of the Cilium metric dashboard. Open the Prometheus sample file in the Cilium community, locate cilium-dashboard.json in the file, and copy it.
2. Log in to Grafana, choose Dashboards, and import the cilium-dashboard.json file.
   

   If you have enabled Interconnect with AOM when installing the Grafana add-on in Step 2: Install Add-ons, change the datasource field in the cilium-dashboard.json file to prometheus-aom, and then import the file.

3. Access the dashboard to view Cilium metrics. For details about Cilium metrics, see cilium-metrics.