Question 20
A database team wants to prevent a recurrence of this issue: Cloud SQL storage grows unexpectedly after an application team enables detailed event logging. Which preventive practice is most relevant?
Explanation
Correct Answer
Include log and audit-table retention in long-range storage forecasts. The original failure arose from: Unbounded database-resident event retention consumes more disk than the sizing model allowed. The control reduces that failure mode.
Incorrect Answers
For the situation in which a Cloud SQL database runs out of free storage during a large index build even though the indexed table itself fits, it makes sense to include maintenance peak-space requirements in database sizing reviews. Here, unbounded database-resident event retention consumes more disk than the sizing model allowed, so the safeguard misses the recurring fault.
For the situation in which a Cloud SQL order database runs out of CPU each afternoon while overnight processing is nearly idle, it makes sense to model overlapping batch and interactive demand rather than average daily utilization. Here, unbounded database-resident event retention consumes more disk than the sizing model allowed, so the safeguard misses the recurring fault.
The proposed check would match geographic failure objectives to Spanner instance configuration rather than assuming all replicas cross regions, helping with cases where a critical ledger on Spanner must tolerate a full regional outage but uses a regional instance configuration. The recurring exposure described here is that unbounded database-resident event retention consumes more disk than the sizing model allowed.