Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Database Engineer practice bank has an additional 769 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your team manages a mission-critical AlloyDB for PostgreSQL production cluster and a matching staging cluster restored from a production backup. Your business requires:
Which maintenance and validation strategy should you implement?
Maintenance timing options such as Week 1 (preview/canary) and Week 2 (production/stable) are scheduling controls used in Cloud SQL and Apigee to stagger rollout waves across weeks after an advance notification is sent. Meanwhile, the maintenance_events log in Logs Explorer records operational lifecycle events indicating when maintenance is scheduled, started, or completed.
Week 1 or Week 2 maintenance timing flags. Instead, AlloyDB controls rollout ordering based on whether a cluster has a preferred maintenance window configured (unscheduled clusters update first, followed at least seven days later by clusters with a scheduled day and hour).maintenance_events in Logs Explorer only confirms that the managed update process started and finished. It does not validate application connectivity, data integrity, query performance, or SLA compliance.Week 1 applies updates 7 to 14 days after notification, Week 2 applies updates 15 to 21 days after notification, and Week 5 applies updates 35 to 42 days after notification.maintenance_events logs to detect when staging has updated.You were managing Cloud SQL instances (rather than AlloyDB) and only needed to determine when maintenance windows were scheduled or completed rather than performing full post-upgrade functional and performance validation.
A staging-first maintenance strategy in AlloyDB for PostgreSQL leverages Google Cloud's update ordering rules to test new database patches, features, and compatibility updates in an identical non-production environment before they reach production.
gcloud alloydb clusters update) aligns automatic restarts with your lowest-traffic period.maintenance_events in Logs Explorer), you compare pre-update baselines with key metrics, execute load tests to detect regressions, analyze query performance, and monitor resource utilization. If issues occur, you configure a deny maintenance period (up to 30 days) on production and open a high-priority support case.Unlike relying on passive status dashboards or arbitrary day-apart schedules, combining an unscheduled staging cluster with a scheduled production window, active load and query testing, and post-maintenance smoke tests satisfies every operational, timing, and verification requirement.
Setting explicit preferred maintenance windows assigns a specific day and UTC hour to each cluster, while Database Center provides a fleet-wide governance dashboard for viewing maintenance schedules, upcoming updates, and historical status across AlloyDB and Cloud SQL resources.
maintenanceUpdatePolicy.maintenanceWindows field must remain unset (or cleared with --maintenance-window-any).Your goal was simply to stagger two independent non-critical clusters across a weekend and audit fleet-wide version status at a high level without performing application or query validation.
--maintenance-window-any and crash recovery doIn AlloyDB for PostgreSQL, the --maintenance-window-any flag clears a cluster's preferred maintenance window, returning it to the default unscheduled state. Crash recovery is an internal database engine mechanism that automatically replays logs upon startup if any incomplete transactions did not finish rolling back during shutdown.
--maintenance-window-any on the production cluster does not postpone or block an upcoming update. Instead, it places production into the unscheduled tier—meaning AlloyDB can update production immediately and at any time, including during peak traffic. To halt a scheduled production update when staging fails, you must configure a deny maintenance period (using --deny-maintenance-period-start-date, --deny-maintenance-period-end-date, and --deny-maintenance-period-time) and open a support case.You intended to convert a scheduled cluster back into an unscheduled staging cluster so it would receive updates as early as possible.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated