Question 320
A Snowflake platform team receives the following requirement: A repeated warehouse scan becomes slower immediately after the warehouse auto-suspends and resumes, even though SQL and data have not changed. Which explanation best accounts for the observed behavior?
Explanation
Correct Answer
The described symptoms indicate that a suspended warehouse loses its local data cache and must warm it again. The appropriate resolution is: Evaluate a suitable auto-suspend interval and monitor data scanned from cache before changing data design. Confirm with: Repeated scans improve after warming while idle compute cost remains within budget.
Incorrect Answers
Economy scaling may tolerate more queuing to conserve compute credits. “Use a scaling policy aligned with queue-latency requirements and measure the tradeoff” would address that alternative cause, not the condition in this question: a suspended warehouse loses its local data cache and must warm it again.
“Consolidate appropriately timed work or choose a suspend policy that reduces needless churn” responds to this separate finding: repeated warehouse resumes can incur minimum billing intervals despite short queries. This question instead needs “Evaluate a suitable auto-suspend interval and monitor data scanned from cache before changing data design”, so the alternative does not meet the stated condition.
Changes to underlying data can invalidate eligibility for persisted result reuse. “Treat result-cache eligibility as conditional and remeasure execution after data changes” applies to that separate problem. The observed case is a repeated warehouse scan becomes slower immediately after the warehouse auto-suspends and resumes, even though SQL and data have not changed; choosing this alternative would address the wrong failure mode.