I usually restart application pods using Argo CD or Rancher. That's convenient for one or two applications, but when there are several deployments to restart, I'd rather do it from the terminal.

First, check the current cluster context and the deployments in your namespace:

kubectl config current-context
kubectl get deployments -n staging

Then trigger a restart for one deployment:

kubectl rollout restart deployment/web -n staging

An example response is:

deployment.apps/web restarted

Replace web and staging with your deployment and namespace. The response confirms that the restart was requested, so I also watch its progress:

kubectl rollout status deployment/web -n staging --timeout=120s

Once complete, you might see:

deployment "web" successfully rolled out

To restart all deployments in that namespace:

kubectl rollout restart deployment -n staging

Or narrow it down using a label that exists on your deployments:

kubectl rollout restart deployment -n staging -l app=web

One distinction worth remembering: this replaces the pods managed by the selected deployments. It doesn't restart every kind of pod in the namespace. Replacement behavior follows each deployment's strategy, so availability still depends on its configuration and available capacity.

You can find more examples in the kubectl rollout reference.

Hope you found this useful!