What is restartPolicy in Kubernetes

restartPolicy tells Kubernetes:

👉 “What should I do when a container inside a Pod stops?”

Kubernetes uses this for self-healing (auto restart).

Types of restartPolicy

There are 3 values only:

1. Always (default)

• Container will restart every time it stops
• Even if it exited successfully ✅

👉 Used for:

• Web apps
• APIs
• Long running services

2. OnFailure

• Restart only if error happens (exit code ≠ 0)
• No restart if success

👉 Used for:

• Batch jobs
• Data processing

3. Never

• Never restart

👉 Used for:

• One-time scripts
• Migration jobs

Where to write restartPolicy

spec:
  restartPolicy: Always | OnFailure | Never

Example 1: restartPolicy = Always

apiVersion: v1
kind: Pod
metadata:
  name: always-pod
spec:
  restartPolicy: Always
  containers:
  - name: nginx
    image: nginx

Command to deploy:

kubectl apply -f always.yaml
kubectl get pods

Behavior:

• Even if container stops → it will restart again and again

Example 2: restartPolicy = OnFailure

apiVersion: v1
kind: Pod
metadata:
  name: fail-pod
spec:
  restartPolicy: OnFailure
  containers:
  - name: test
    image: busybox
    command: ["sh", "-c", "exit 1"]

Command:

kubectl apply -f onfailure.yaml
kubectl get pods

Behavior:

• exit 1 → restart
• exit 0 → no restart

Example 3: restartPolicy = Never

apiVersion: v1
kind: Pod
metadata:
  name: never-pod
spec:
  restartPolicy: Never
  containers:
  - name: test
    image: busybox
    command: ["sh", "-c", "echo Hello && exit 0"]

Command:

kubectl apply -f never.yaml
kubectl get pods

Behavior:

• Runs once → stops → no restart

Quick kubectl command

-- Always
kubectl run test --image=nginx --restart=Always
-- OnFailure
kubectl run test --image=busybox --restart=OnFailure -- /bin/sh -c "exit 1"
-- Never
kubectl run test --image=busybox --restart=Never -- echo Hello

How to check restartPolicy

kubectl get pod <pod-name> -o yaml | grep restartPolicy