A Snowflake administrator is configuring a virtual warehouse for a financial reporting workload. The warehouse is currently sized at LARGE. During month-end close, query queue times increase significantly, but the team does not want to pay for a larger warehouse during the rest of the month. The administrator decides to enable multi-cluster warehouses. Which combination of settings will allow Snowflake to add clusters ONLY when queries are queuing AND remove extra clusters as soon as the queue clears, while keeping costs as low as possible during normal periods?
- A. MIN_CLUSTER_COUNT = 1, MAX_CLUSTER_COUNT = 4, SCALING_POLICY = 'ECONOMY'✓ Correct
- B. MIN_CLUSTER_COUNT = 1, MAX_CLUSTER_COUNT = 4, SCALING_POLICY = 'STANDARD'
- C. MIN_CLUSTER_COUNT = 2, MAX_CLUSTER_COUNT = 4, SCALING_POLICY = 'STANDARD'
- D. MIN_CLUSTER_COUNT = 1, MAX_CLUSTER_COUNT = 1, SCALING_POLICY = 'ECONOMY'
Explanation
CORRECT — MIN_CLUSTER_COUNT = 1 ensures only one cluster runs during normal periods (minimizing cost), MAX_CLUSTER_COUNT = 4 allows scale-out during peak demand, and SCALING_POLICY = 'ECONOMY' waits until a cluster is fully loaded before spinning up a new one AND shuts down extra clusters as soon as their queue clears, prioritizing cost savings over speed. OPTION B is wrong because STANDARD policy scales up more aggressively (before queuing becomes severe) and is slower to scale down, leading to higher costs during the transition periods — it does not minimize cost as effectively as ECONOMY. OPTION C is wrong because MIN_CLUSTER_COUNT = 2 guarantees two clusters run at all times, incurring unnecessary cost during normal (non-peak) periods. OPTION D is wrong because MAX_CLUSTER_COUNT = 1 disables scale-out entirely; this is effectively just a single-cluster warehouse and will not reduce queue times during peak load regardless of the scaling policy.