Kubernetes Liveness, Readiness Probe Explained

In Kubernetes, other than the restartPolicy which defaults to restart the pod when a pod fails. There are Liveness and Readiness probes to control and manage the lifecycle of a pod.

Liveness Probe

A liveness Probe is used by the kubelet to know when to restart the container inside a pod. This probe can help restart the container when the application is not responding to the request or the response time is too long.

There are 3 main ways to check the container’s health using a liveness probe.

1. Liveness command

apiVersion: v1
kind: Pod
metadata:
name: webserver
spec:
containers:
  - name: nginx
    image: nginx:latest
    livenessProbe:
        exec:
            command:
            - cat
            - /tmp/healthy
        initialDelaySeconds: 3
        periodSeconds: 30

We are creating a nginx container inside the pod webserver. The container has a livenessProbe attached to it.

exec:
    command:
    - cat
    - /tmp/healthy

If the command under exec fails the kubelet will restart the container. In this case, if there is no file at /tmp/healthy the container will restart.

initialDelaySeconds: 3

The value initialDelaySeconds tells the kubelet to wait 3 seconds before performing the first probe. This is used to delay the probe when the container/application is starting up. Default is 0

periodSeconds: 30

The value periodSeconds is the time between each probe. Default is 10

2. Liveness HTTP request

livenessProbe:
    httpGet:
        path: /healthz
        port: 8080
    initialDelaySeconds: 3
    periodSeconds: 30

Replace the exec under livenessProbe with httpGet.
The kubelet will send a GET request to the /healthz endpoint on port 8080 of the container. If the handler returns a success code, the kubelet considers the container to be healthy. If the handler returns a failure code the kubelet will restart the container