Question

A development team reports that a newly deployed pod on a private GKE cluster is stuck in CrashLoopBackOff. As the cloud architect, you need to guide an on-call engineer who only has access to a browser to inspect the pod's container logs and recent events without provisioning any local tools or bastion host. Which approach should you advise?

Accepted Answer

Open Cloud Shell, run 'gcloud container clusters get-credentials' to fetch cluster credentials, then use 'kubectl logs' and 'kubectl describe pod' to inspect the failing pod.

Answer

Instruct the engineer to install the gcloud CLI and kubectl on their laptop, configure kubeconfig manually, and connect over the public endpoint.

Answer

SSH directly into the underlying GKE node VM using the Compute Engine console and read the container runtime logs from the node's local filesystem.

Answer

Enable Serial console access on the node pool and read the pod logs from the VM serial output stream in the console.