Liveness Probes

All command is same as Readiness Probes, but the only difference is the livenessProbe field.

Usage of Liveness Probes

The concept of Liveness probes basically is to determine whether a container is alive or not by performing some tests. If the liveness probes are successful, then we know that container is going to run (healthy), else, it (kubelet) will restart the container (unhealthy).

Let me give you one scenario, suppose you have a container running an API server, and the API server is not responding to the requests, but the container still stay alive. In this case, the liveness probe will help you to restart the container, so that a new container can be created and the API server can start responding to the requests.

HTTP

liveness-probes.yaml


apiVersion: v1
kind: Pod
metadata:
  name: sample-pod
spec:
  containers:
    - name: web-app
      image: webapp
      livenessProbe:
        httpGet:
          path: /api/v1/health
          port: 8000
        initialDelaySeconds: 5 # Number of seconds after the container has started before startup, that means it should wait for 5 seconds before performing the 1st probe
        periodSeconds: 3 # how often to perform the probe
        failureThreshold: 5 # default 3, if the application is not ready after 3 attempts, then the probe will stop

TCP

liveness-probes.yaml


apiVersion: v1
kind: Pod
metadata:
  name: sample-pod
spec:
  containers:
    - name: web-app
      image: webapp
      livenessProbe:
        tcpSocket:
          port: 8080
        initialDelaySeconds: 5 # Number of seconds after the container has started before startup, that means it should wait for 5 seconds before performing the 1st probe
        periodSeconds: 3 # how often to perform the probe
        failureThreshold: 5 # default 3, if the application is not ready after 3 attempts, then the probe will stop

kubelet will try to open a socket to the container on the specified port 8080. If it can establish the connection, then the state is healthy.

exec

liveness-probes.yaml


apiVersion: v1
kind: Pod
metadata:
  name: sample-pod
spec:
  containers:
    - name: web-app
      image: webapp
      livenessProbe:
        exec:
          command:
            - cat
            - /app/healthy
        initialDelaySeconds: 5 # Number of seconds after the container has started before startup, that means it should wait for 5 seconds before performing the 1st probe
        periodSeconds: 3 # how often to perform the probe
        failureThreshold: 5 # default 3, if the application is not ready after 3 attempts, then the probe will stop

It will return 0 status code if the command succeed.