Question 340
The service owner asks how to avoid a repeat of: Cloud SQL request latency spikes during lunchtime although average CPU remains normal. Which operational check is the best match?
Explanation
Correct Answer
Alert on long transactions and lock-wait duration before they degrade service. The recurring risk is slow queries are waiting on row locks held by long-running update transactions. This check makes that risk visible before another incident.
Incorrect Answers
The proposed check would define and observe data freshness SLOs for every replica-based report, helping with cases where read replica utilization remains modest but users see stale dashboards after daily imports. The recurring exposure described here is that slow queries are waiting on row locks held by long-running update transactions.
The proposed check would document legal-hold preservation separately from standard backup schedules, helping with cases where an e-commerce database has backups stored for 30 days while a legal hold requires preserving selected records longer. The recurring exposure described here is that slow queries are waiting on row locks held by long-running update transactions.
This control would track memory-sensitive query patterns and concurrent execution costs; it responds to the risk in which a service receives intermittent out-of-memory query failures during complex sorting operations. It would not address the current root cause: slow queries are waiting on row locks held by long-running update transactions.