Professional Cloud DevOps Engineer Practice Test

What this test covers

One concept per question, across the topics Professional Cloud DevOps Engineer draws on.

• Canary releases validate against real traffic before full rollout
• Workload Identity removes long-lived credentials from the cluster
• An error budget is a pre-agreed policy that governs release decisions
• Build once, promote the same artefact through every environment
• Post-mortems improve systems, not accountability records
• Pending pods indicate a scheduling problem, almost always insufficient capacity
• Traces show where time is spent across service boundaries
• Log-based metrics turn log content into numeric time series you can alert on
• Safe rollouts require readiness signalling and graceful connection draining
• Alert on symptoms users experience, not on causes that may be harmless
• Burn-rate alerting scales urgency to how quickly the budget is being consumed
• Cloud Run concurrency directly determines how many instances a given request rate needs
• Autopilot removes node management in exchange for reduced configurability
• Namespaces alone provide no network isolation; network policies do
• Baking a custom image minimises startup time and makes instances reproducible
• The four golden signals describe service health from the user's perspective
• Pod scaling and node scaling are separate mechanisms
• Enforce at both build and admission so the control cannot be bypassed