Question 40
A postincident review documents: Bigtable CPU remains uneven after a team doubles node capacity for a time-series workload. Which ongoing check targets the underlying failure?
Explanation
Correct Answer
Test row-key and prefix distribution before scaling a Bigtable ingestion pipeline. The original failure arose from: The schema directs most reads and writes to a narrow key prefix so adding nodes cannot remove the hotspot. The control reduces that failure mode.
Incorrect Answers
The stated control responds to a failure mode where a payroll database is moved to the cheapest Cloud SQL configuration and batch reports take far longer than before. It does not prevent this one: the schema directs most reads and writes to a narrow key prefix so adding nodes cannot remove the hotspot.
The proposed check would include patching, backups, monitoring, and staffing in database TCO analyses, helping with cases where a small team installs and patches its own database on virtual machines despite having no custom engine requirements. The recurring exposure described here is that the schema directs most reads and writes to a narrow key prefix so adding nodes cannot remove the hotspot.
This would test reporting workloads against OLTP latency budgets before onboarding them. It is useful when a Cloud SQL application performs large analytical joins that stall small OLTP transactions on the same database, but it does not prevent the issue here: the schema directs most reads and writes to a narrow key prefix so adding nodes cannot remove the hotspot.