kube-apiserver Dashboard

Home > Select Cluster Project > Cluster > kube-apiserver Dashboard

Note

Requirements

• The Kubernetes cluster project (CP) requires the Read role.

• The WhaTap Kubernetes agent 1.5.6 or later is required.

The kube-apiserver is a component that handles the requests inside and outside the Control Plane. After collecting the status and performance metric of kube-apiserver through the kube-apiserver dashboard menu, you can easily monitor the availability and performance of the cluster by tracing the kube-apiserver's response time, throughput, request failure, and such.

Basic screen guide

The kube-apiserver dashboard menu reflects kube-apiserver's key metrics for the last one hour at an interval of 1 minute. You can see the following metrics. For more information about metrics, see the Metric guide below.

• To monitor the cluster's load and usage, the status and trends for request counts by method, resource, and response code.

• The number of API requests being processed is provided to help you understand the cluster load.

• It provides the number of created OS threads and current goroutines.

Enabling the kube-apiserver monitoring

When installing the WhaTap Kubernetes agent, modify the yaml file downloaded according to the following instructions to enable the kube-apiserver monitoring.

1. Add the agent container specifications for kube-apiserver monitoring under the spec of the whatap-master-agent deployment.

containers:
  - name: whatap-control-plane-helper
    image: whatap/kube_mon
    imagePullPolicy: Always
    ports:
      - containerPort: 9496
    command: ["/data/agent/master/whatap_control_plane_helper"]
    args: ["-collect_control_plane_monitoring_enabled=true"]

See the following example:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: whatap-master-agent
  namespace: whatap-monitoring
spec:
  replicas: 1
  selector:
    matchLabels:
      name: whatap-master-agent
  template:
    metadata:
      labels:
        name: whatap-master-agent
    spec:
      containers:
        - name: whatap-control-plane-helper
          image: whatap/kube_mon
          imagePullPolicy: Always
          ports:
            - containerPort: 9496
          command: ["/data/agent/master/whatap_control_plane_helper"]
          args: ["-collect_control_plane_monitoring_enabled=true"]
        - name: whatap-master-agent
          image: whatap/kube_mon
          imagePullPolicy: Always
          resources:
            requests:
              memory: 300Mi
              cpu: 100m
            limits:
              memory: 350Mi
              cpu: 200m
          ports:
            - containerPort: 6600
          command: ['/bin/entrypoint.sh']
          env:
            - name: WHATAP_LICENSE
              value: "xxxxxxxx-12345xxxxxx-x00e000000xxx8"
            - name: WHATAP_HOST
              value: "13.124.11.223/13.209.172.35"
            - name: WHATAP_PORT
              value: "6600"
            - name: WHATP_MEM_LIMIT
              valueFrom:
                resourceFieldRef:
                  containerName: whatap-master-agent
                  resource: limits.memory
          volumeMounts:
            - name: start-script-volume
              mountPath: /bin/entrypoint.sh
              readOnly: true
              subPath: entrypoint.sh
            - mountPath: /whatap_conf
              name: whatap-config-volume
      volumes:
        - name: start-script-volume
          configMap:
            defaultMode: 0700
            name: master-start-script
        - name: whatap-config-volume
          emptyDir: {}
      serviceAccount: whatap

1. Add the following to the env settings in the whatap-master-agent container specifications of the whatap-master-agent deployment.

env:
  - name: collect_control_plane_monitoring_enabled
    value: "true"

See the following example:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: whatap-master-agent
  namespace: whatap-monitoring
spec:
  replicas: 1
  selector:
    matchLabels:
      name: whatap-master-agent
  template:
    metadata:
      labels:
        name: whatap-master-agent
    spec:
      containers:
        - name: whatap-control-plane-helper
          image: whatap/kube_mon
          imagePullPolicy: Always
          ports:
            - containerPort: 9496
          command: ["/data/agent/master/whatap_control_plane_helper"]
          args: ["-collect_control_plane_monitoring_enabled=true"]
        - name: whatap-master-agent
          image: whatap/kube_mon
          imagePullPolicy: Always
          resources:
            requests:
              memory: 300Mi
              cpu: 100m
            limits:
              memory: 350Mi
              cpu: 200m
          ports:
            - containerPort: 6600
          command: ['/bin/entrypoint.sh']
          env:
            - name: collect_control_plane_monitoring_enabled
              value: "true"
            - name: WHATAP_LICENSE
              value: "xxxxxxxx-12345xxxxxx-x00e000000xxx8"
            - name: WHATAP_HOST
              value: "13.124.11.223/13.209.172.35"
            - name: WHATAP_PORT
              value: "6600"
            - name: WHATP_MEM_LIMIT
              valueFrom:
                resourceFieldRef:
                  containerName: whatap-master-agent
                  resource: limits.memory
          volumeMounts:
            - name: start-script-volume
              mountPath: /bin/entrypoint.sh
              readOnly: true
              subPath: entrypoint.sh
            - mountPath: /whatap_conf
              name: whatap-config-volume
      volumes:
        - name: start-script-volume
          configMap:
            defaultMode: 0700
            name: master-start-script
        - name: whatap-config-volume
          emptyDir: {}
      serviceAccount: whatap

1. Add the following roles to the whatap ClusterRole to collect the control plane monitoring performance metrics.

rules:
  - nonResourceURLs: ["/metrics"]
    verbs: ["*"]

See the following example:

apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  annotations:
    rbac.authorization.kubernetes.io/autoupdate: "true"
  name: whatap
  namespace: whatap-monitoring
rules:
  - apiGroups: ["*"]
    resources: ["namespaces", "pods", "pods/log", "configmaps","services", "endpoints","daemonsets","ingresses","deployments","nodes","persistentvolumes","persistentvolumeclaims", "events", "replicasets", "roles", "rolebindings", "clusterroles", "clusterrolebindings", "jobs", "cronjobs", "statefulsets", "serviceaccounts", "configmaps", "storageclasses", "horizontalpodautoscalers" ]
    verbs: ["get","list","watch"]
  - apiGroups: [""]
    resources: ["pods/exec"]
    verbs: ["create"]
  - apiGroups: [""]
    resources: ["configmaps"]
    verbs: ["*"]
  - nonResourceURLs: ["/metrics"]
    verbs: ["*"]
---

Note

You can also enable the kube-apiserver monitoring via the kubectl apply or kubectl edit command.

kube-apiserver metric

Classification	Metric	Description
API GET request	GET requests per minute	Number of GET requests per minute of kube-apiserver
	Trend of GET request counts per second	Trend of GET request counts per second of kube-apiserver
	Trend of GET request counts per minute for each resource and subresource	Trend of the GET request counts per minute by resource and subresource incoming to kube-apiserver
API POST request	POST requests per minute	Number of POST requests per minute of kube-apiserver
	Trend of POST request counts per second	Trend of POST request counts per second of kube-apiserver
	Trend of POST request counts per minute for each resource and subresource	Trend of the POST request counts per minute by resource and subresource incoming to kube-apiserver
Requests by API response code	[ 2xx ] Request count per minute	The number of requests per minute with the response codes between 200 and 300 among kube-apiserver requests.
	[ 2xx ] Trend of request counts per minute for each response code	Response code-based request counts per minute with the response codes between 200 and 300 among kube-apiserver requests.
	[ 2xx ] Trend of request counts per minute for each resource and subresource	Trend of request counts per minute grouped by resource and subresource with the response codes between 200 and 300 among kube-apiserver requests.
	[ 3xx ] Request count per minute	The number of requests per minute with the response codes between 300 and 400 among kube-apiserver requests.
	[ 3xx ] Trend of request counts per minute for each response code	Response code-based request counts per minute with the response codes between 300 and 400 among kube-apiserver requests.
	[ 3xx ] Trend of request counts per minute for each resource and subresource	Trend of request counts per minute grouped by resource and subresource with the response codes between 300 and 400 among kube-apiserver requests.
	[ 400~ ] Request count per minute	The number of requests per minute with the response codes above 400 among kube-apiserver requests.
	[ 400~ ] Trend of request counts per minute for each response code	Response code-based request counts per minute with the response codes below 400 among kube-apiserver requests.
	[ 400~ ] Trend of request counts per minute for each resource and subresource	Trend of request counts per minute grouped by resource and subresource with the response codes above 400 among kube-apiserver requests.
Go Metrics	Number of goroutines	Number of current goroutines
	Number of threads	Current OS thread count
Number of active requests (1 second)	Number of requests run in the last 1 second	Number of requests run within 1 second in kube-apiserver during data collection