Add Readiness Probe

Original Prepium workshop · Tested on Kubernetes v1.35.8 on 8 September 2026; NetworkPolicy checks used Calico v3.32.2. Run this independent example in a disposable cluster. It does not require the simulator's private task files.

A running container can still be unready

The web server below serves / on port 80. Its readiness probe requests a path that does not exist. The container starts, but the Pod stays unready and the Service should have no ready backend.

1. Create the failing workload

Create prepium-tutorial if it does not already exist. Save this as probe-broken.yaml and apply it.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: probe-web
  namespace: prepium-tutorial
spec:
  replicas: 1
  selector:
    matchLabels:
      app: probe-web
  template:
    metadata:
      labels:
        app: probe-web
    spec:
      containers:
        - name: web
          image: nginx:1.27-alpine
          ports:
            - containerPort: 80
          readinessProbe:
            httpGet:
              path: /not-a-health-endpoint
              port: 80
            periodSeconds: 3
---
apiVersion: v1
kind: Service
metadata:
  name: probe-web
  namespace: prepium-tutorial
spec:
  selector:
    app: probe-web
  ports:
    - port: 80
      targetPort: 80

kubectl create namespace prepium-tutorial
kubectl apply -f probe-broken.yaml
kubectl -n prepium-tutorial get pods -l app=probe-web
kubectl -n prepium-tutorial describe pods -l app=probe-web
kubectl -n prepium-tutorial get endpointslices -l kubernetes.io/service-name=probe-web -o yaml

Look for 0/1 ready containers, an HTTP 404 probe failure in events, and ready: false on the EndpointSlice endpoint. A Pod can be in phase Running while failing readiness; the phase alone is insufficient evidence.

2. Repair the application check

For this stock nginx workshop, / is a real served path. Change readinessProbe.httpGet.path to / in the file, keep port 80, and apply again. For your own application, choose an endpoint that actually represents readiness to serve traffic.

kubectl apply -f probe-broken.yaml
kubectl -n prepium-tutorial rollout status deployment/probe-web --timeout=120s
kubectl -n prepium-tutorial wait --for=condition=Ready pod -l app=probe-web --timeout=120s
kubectl -n prepium-tutorial get endpointslices -l kubernetes.io/service-name=probe-web -o yaml

Now the new Pod should be ready and its EndpointSlice endpoint should show ready: true. The Deployment change creates a new Pod; check the current ReplicaSet rather than reading old events from a terminating Pod.

3. Test the other common mistake

Change the probe's port to 8080, while leaving the server on 80. Expect a connection failure instead of an HTTP 404. Increasing the timeout cannot correct a wrong port. Restore port 80 and repeat the readiness and EndpointSlice checks.

Deleting the readiness probe would hide this particular failure without proving that the probe requirement was satisfied. In a scored task, both the configured check and the resulting ready state may matter. Read the actual requirements before making the workload appear green.

Readiness controls readiness for traffic; a failed liveness check can restart a container. Do not interchange them simply to make a test pass. See Kubernetes probe configuration.

Cleanup: kubectl delete namespace prepium-tutorial removes this workshop. Next, try Service selectors and rollout verification.