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.
A team runs a PostgreSQL database on Cloud SQL in us-central1 as a standalone instance in us-central1-a with a read replica in us-central1-b for reporting. During a simulated outage in us-central1-a, automatic failover did not occur, recent commits were lost during manual replica promotion, and applications failed due to an IP address change.
The redesigned architecture must meet these requirements:
Which configuration should the team implement?
us-central1-a and us-central1-b, and deploy a read replica in us-central1-cA Cloud SQL instance configured for high availability (HA) operates as a regional instance consisting of a primary instance in one zone (us-central1-a) and a standby instance in a secondary zone (us-central1-b) within the same region. Pairing this setup with a read replica in a third zone (us-central1-c) isolates read scaling from primary zone failures.
us-central1-a and us-central1-b before a transaction is reported as committed, guaranteeing zero data loss (RPO = 0) during a zonal outage.us-central1-c ensures reporting queries continue uninterrupted if us-central1-a goes down. Once failover finishes, the replica automatically connects to the newly promoted primary in us-central1-b.us-central1-a.Unlike standalone deployments relying on asynchronous read-replica promotion or manual Point-in-Time Recovery, regional HA automates failover with zero data loss and preserves the static IP, while a third-zone read replica properly handles reporting traffic that the passive standby cannot serve.
us-central1-a with automated backups and point-in-time recovery enabled, and deploy a read replica in us-central1-cAutomated backups combined with Point-in-Time Recovery (PITR) store database snapshots and transaction logs in Cloud Storage, enabling administrators to restore a database to a specific moment in time—typically to recover from accidental deletions or logical data corruption.
DROP TABLE commands) rather than providing continuous infrastructure availability.us-central1-c survives a us-central1-a outage, write operations remain completely unavailable until manual restoration completes.The primary goal were protecting a non-critical database against logical data corruption or accidental modifications, where manual recovery, up to five minutes of data loss, and updating client connection strings were acceptable.
us-central1-a and us-central1-b, and route reporting queries directly to the standby instanceIn a Cloud SQL regional high availability (HA) deployment, the standby instance serves exclusively as a passive failover target in a secondary zone. It stays synchronized with the primary instance through synchronous replication to regional persistent disks, ready to take over if the primary instance or zone becomes unresponsive.
While this configuration provides synchronous cross-zone replication and automatic failover with a shared static IP address, it fails the read-offloading requirement because a Cloud SQL standby instance cannot be used for read queries. Any attempt to point reporting workloads directly at the standby fails because no read endpoint is exposed on the passive standby node.
us-central1-c) to remain isolated from outages in both the primary and standby zones.The workload were deployed on AlloyDB Omni using its Enhanced Availability architecture, where standby read-replicas actively serve read-only reporting traffic while also providing failover redundancy, or if the team had no requirement to offload reporting queries from the primary instance.
us-central1-a, deploy read replicas in us-central1-b and us-central1-c, and rely on automatic replica promotion to keep the static IPRead replicas are asynchronous copies of a Cloud SQL primary instance designed to offload read-heavy workloads—such as reporting and analytics—from the primary database. They can be deployed across different zones within a region to improve read availability.
The team only needed horizontal read scalability across multiple zones and could tolerate manual intervention, potential data loss from replication lag, and updating application connection strings after a zonal failure.
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