Quick revision for the Associate Cloud Engineer exam.

Treat projects as operational boundaries for APIs, quotas, billing, IAM, and resources.

Choose Compute Engine for OS control, Cloud Run for managed containers, and GKE when Kubernetes is a requirement.

Use managed instance groups for repeatable, autoscaled, self-healing VM capacity.

Match object, block, and file interfaces to Cloud Storage, persistent disks or Hyperdisk, and Filestore.

Separate OLTP, analytics, documents, wide-column, and cache requirements before choosing a database.

Trace traffic through IP addressing, routes, firewall policy, DNS, NAT, private access, and load balancing.

Start troubleshooting with inventory and evidence: status, events, metrics, logs, traces, and audit records.

Separate human identities, service accounts, workload federation, and runtime identities.

Prefer least privilege, short-lived credentials, reproducible changes, and tested backup or recovery paths.