4.1.1. Compute Engine Operations
💡 First Principle: Compute Engine's day-to-day operational surface splits cleanly into three activities — connecting to what's running, seeing what's running, and protecting what's running via backups — and the exam tests whether you can name the right tool or view for each.
Remotely connecting to an instance typically means SSH (via the console's built-in SSH-in-browser, the gcloud compute ssh command, or a third-party client using OS Login-managed keys). Viewing currently running instances is a console or gcloud compute instances list task, filterable by zone, label, or status. Snapshots capture an incremental, point-in-time copy of a Persistent Disk's data — useful for backup and disaster recovery — and can be scheduled to run automatically on a recurring basis rather than requiring a manual trigger every time. Images, by contrast, are boot-disk templates: a fixed, versioned OS-plus-software configuration used to create new VM instances consistently, whether that's scaling out identical web servers or standardizing a golden image across a fleet.
| Concept | Purpose | Typical Use |
|---|---|---|
| Snapshot | Incremental backup of disk data | Disaster recovery, point-in-time restore |
| Image | Boot-disk template | Creating new, identically-configured VMs |
| Scheduled snapshot | Automated recurring snapshot | Ongoing backup policy without manual triggers |
⚠️ Exam Trap: A scenario asking how to guarantee a fleet of new VMs all launch with an identical, pre-configured software stack is asking about images, not snapshots — snapshots protect existing disk data, they aren't the mechanism for standardizing new instance creation.
Reflection Question: A team wants both an automated nightly backup of a production disk's data and a way to spin up new identically-configured instances on demand. Which two distinct features does that require, and why doesn't one alone cover both needs?
