GCP Associate Cloud Engineer Practice Question 87: Cloud SQL Read Replica Reporting
Google Cloud Associate Cloud Engineer · Test 9 · question 7 of 10
A production Cloud SQL for PostgreSQL instance serves a transactional web application. A new internal reporting dashboard needs to run large, frequent SELECT queries against the same data, but the team is worried the extra query load will slow down the primary instance for the application's users. The reporting queries never need to write data. What should they configure to isolate the reporting load from the primary instance?
Show answer and explanation
Answer:
A Cloud SQL read replica is a separate, independently scalable copy of the instance that stays in sync with the primary through replication and accepts read-only queries. Pointing the reporting workload at a replica offloads its read traffic entirely, so it cannot compete with the application's transactional queries on the primary.
Option B is a different feature entirely; the standby in a high availability configuration is reserved for automatic failover and is not intended to serve regular read traffic. Option C throws more compute at both workloads on the same instance, but the two workloads still contend for the same resources and connections, and it costs more than a right-sized replica. Option D does not exist as described; automated backups are for recovery, not a queryable live endpoint, and reporting against a backup snapshot would mean stale, static data rather than current information.
Reference: About replication in Cloud SQL