A GKE Deployment named echo-deployment (exposed by echo-service) needs an application update with minimal downtime. What should you do?

Using kubectl set image deployment/echo-deployment triggers a rolling update by default for Deployments. This ensures new pods with the updated image are created and become ready before old pods are terminated, minimizing downtime. The rolling update functionality of the instance group is for VM instances, not for individual application containers within Kubernetes. Deleting and recreating the deployment would cause downtime because all old pods are terminated before new ones are created. Updating the service YAML with the new image is incorrect because services define how to access pods, not the application image itself.