Google ACE Practice Question: A financial application requires all database…

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Cloud SQL read replica with synchronous replication

  Why it's wrong here

  Cloud SQL read replicas always use asynchronous replication from the primary to the replica, which means there is inherent replication lag—the replica may be missing recent transactions if the primary fails. Because of that lag, a read replica cannot guarantee zero data loss, so the RPO is not near-zero. Read replicas are intended to offload read traffic and scale query throughput, not to provide high availability; Cloud SQL does not offer a synchronous replication mode for read replicas.

• ✓

  Cloud SQL High Availability with synchronous standby replica in a different zone

  Why this is correct

  Cloud SQL High Availability (HA) provisions a primary and a standby instance in different zones within the same region. It uses synchronous replication to the standby, so every committed transaction is durably written to both the primary and standby before the client receives an acknowledgement, yielding near-zero RPO. If the primary zone experiences an outage, the standby is promoted automatically and, because replication was synchronous, no committed data is lost even in a hard failure scenario.

• ✗

  Cloud SQL with automated daily backups and point-in-time recovery

  Why it's wrong here

  Automated daily backups and point-in-time recovery (PITR) restore from backup files and write-ahead logs, but backups are taken only once per day and transaction logs are captured asynchronously. The recovery point objective (RPO) therefore is bounded by the time since the last backup or the last log flush—potentially minutes or hours—not the near-zero RPO of a synchronous standby. This approach is designed for disaster recovery against corruption or accidental deletion, not for instantaneous zone failure with zero data loss.

• ✗

  Cloud Spanner — it automatically replicates synchronously across zones

  Why it's wrong here

  Cloud Spanner is a fully managed relational database that offers global strong consistency and synchronous replication across zones and regions, so it could certainly satisfy near-zero RPO requirements. However, this question specifically asks for a Cloud SQL feature; Spanner is a separate product, not a configuration of Cloud SQL. The correct cloud-native answer within the Cloud SQL family is HA with a synchronous standby, so Spanner is technically capable but out of scope for the intended product category.

Go deeper

Related to this question