Google ACE Practice Question: Running a Cloud SQL for MySQL instance that…

Answer analysis

Option-by-option breakdown

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

• ✗

  Enable automatic storage increase on the primary instance

  Why it's wrong here

  Enabling automatic storage increase only expands the disk when capacity is nearly exhausted; it does nothing to affect the CPU or memory pressure caused by read-heavy traffic. The primary instance still executes every read and write query, so the load that is saturating the instance remains completely unchanged. This option addresses a storage-full scenario, not a performance bottleneck from excessive reads.

• ✗

  Increase the machine type of the primary instance

  Why it's wrong here

  Increasing the machine type vertically scales the primary instance by adding more vCPUs and memory, which may help the instance cope with more concurrent queries, but it does not offload any of the read traffic. Every SELECT still hits the same primary node and competes with writes for the same resources. This is a temporary band-aid that does not reduce the root cause—a read-heavy query distribution—and it becomes expensive and unsustainable as traffic grows.

• ✓

  Create one or more read replicas and direct read queries to them

  Why this is correct

  Creating one or more read replicas and directing read-only queries to them is the correct architectural fix for offloading read traffic from the primary instance. Cloud SQL for MySQL read replicas are read-only copies that use asynchronous replication, so they can serve SELECTs without contending with the primary's write operations. This reduces the primary's CPU utilization by distributing the query workload across additional instances, and it is a standard pattern for scaling read-heavy applications.

• ✗

  Change the instance to use private IP only

  Why it's wrong here

  Changing the instance to use private IP only removes public internet exposure and routes traffic through the VPC, which improves security and may reduce network overhead, but it does not change the fact that the primary instance still processes all queries. CPU consumption on the primary is driven by query execution, not by the network path the queries take. Since the option does not alter the read-to-write distribution or add compute capacity, it has no meaningful effect on the high CPU problem.

Go deeper

Related to this question