An instance in a Compute Engine group is causing failures and is unresponsive to SSH. What is the next diagnostic step?

When a VM is unresponsive to SSH, inspecting the serial-port output is the most effective diagnostic step. The serial console provides low-level boot messages, kernel panics, and other system output that can reveal the root cause of the unresponsiveness, even if the OS isn't fully booted or networking isn't functional. Rebooting the VM might temporarily resolve the issue but doesn't provide insight into the problem. Deleting and replacing the VM is a destructive action that loses diagnostic information. Snapshotting the disk is useful for data recovery or deeper forensic analysis, but it's a more involved process and doesn't offer real-time diagnostic output like the serial console.