Replica count

Learn the Pod replica count with the replicas field.

Replica count

The replica count changes in two ways: editing replicas in the YAML and applying again (declarative) or kubectl scale (imperative). The declarative way is preferred; otherwise the next apply undoes your manual change:

kubectl scale deployment web --replicas=5
# set replicas: 5 in web.yaml and:
kubectl apply -f web.yaml

Setting the replica count to 0 stops the application completely without deleting it.

Watching scaling

After scaling, kubectl rollout status waits until all replicas are ready; it is used in CI/CD scripts to make sure the deployment has finished:

$ kubectl scale deploy web --replicas=0
$ kubectl get deploy web
NAME   READY   UP-TO-DATE   AVAILABLE
web    0/0     0            0
$ kubectl scale deploy web --replicas=3
$ kubectl rollout status deploy/web
deployment "web" successfully rolled out

Exercise Stop and start again

Write three commands: scale the web deployment down to 0 replicas; then up to 4; wait for the rollout to finish.

Show example solution

kubectl scale deployment web --replicas=0
kubectl scale deployment web --replicas=4
kubectl rollout status deployment/web

Test yourself

You scaled to 5 with kubectl scale, but web.yaml says replicas: 3. How many replicas after kubectl apply -f web.yaml?
- 0
- 3
- 5
- 8
Show answer

3 — The file is the desired state; apply applies it again.

Wait until all replicas are ready.

kubectl rollout ____ deploy/web

Show answer

kubectl rollout status deploy/web

Match.

kubectl scale

Imperative scaling
replicas: N + apply

Declarative scaling
--replicas=0

Stop without deleting
rollout status

Wait until done

Next lesson: Rolling update

We Are Enhancing Your Experience

Cookie Preferences

Strictly Necessary Cookies

Analytical Cookies