ADA-C01 Question 381
Single answer4.5 Manage and optimize costs.A Snowflake administrator is reviewing monthly spend and finds that compute costs are higher than expected for an ad hoc analytics workload. The workload is run by business analysts during business hours, and query history shows many short queries with long idle gaps between bursts of activity. Analysts have also been leaving worksheets open, causing warehouses to remain available even when no queries are running. The team wants to reduce compute cost without significantly affecting user experience. Which action should the administrator take?
- A
Configure the warehouse with AUTO_SUSPEND set to a low value such as 60 seconds and keep AUTO_RESUME enabled
- B
Convert the warehouse to a multi-cluster warehouse with MAX_CLUSTER_COUNT set to 10 so it can shut down faster during idle periods
- C
Disable query result caching so repeated analyst queries do not return cached results and warehouse usage is easier to measure
- D
Increase the warehouse size from MEDIUM to LARGE so short queries complete faster and total monthly credits decrease
Show answer and explanation
Correct answer: A
Explanation
This scenario focuses on managing and optimizing Snowflake compute costs for intermittent, interactive workloads. Snowflake virtual warehouses consume credits while running, including during idle periods until they suspend. For analyst-driven workloads with many short queries and long pauses, the most effective action is usually to reduce idle time by setting an appropriately low AUTO_SUSPEND value and enabling AUTO_RESUME. This preserves usability while minimizing unnecessary warehouse runtime. Multi-cluster warehouses are better suited for concurrency scaling, not idle-cost reduction. Result caching should generally remain enabled because it reduces compute usage. Increasing warehouse size can help performance in some cases, but it is not the best response when the primary waste is idle warehouse time. These recommendations align with Snowflake best practices for warehouse management, auto-suspension/auto-resume configuration, and compute cost optimization.
- A. Correct.
Correct. For workloads with intermittent usage and frequent idle gaps, setting a low AUTO_SUSPEND value reduces the time a warehouse runs while idle, directly lowering compute charges. Keeping AUTO_RESUME enabled preserves the user experience because the warehouse automatically restarts when a new query arrives. This is a standard Snowflake cost-optimization best practice for interactive, bursty workloads.
- B. Incorrect.
Incorrect. Multi-cluster warehouses are primarily used to improve concurrency handling and can increase compute consumption if additional clusters start. They do not exist to make warehouses suspend faster during idle periods. In this scenario, the problem is idle time between bursts, not sustained concurrency pressure.
- C. Incorrect.
Incorrect. Disabling query result caching would typically increase compute cost, because repeated queries that could have been served from cache would instead re-execute and consume warehouse credits. Easier measurement is not a valid cost-optimization reason to turn off caching for analysts.
- D. Incorrect.
Incorrect. Increasing warehouse size may reduce execution time for some queries, but it does not address the root cause here: warehouses staying running during long idle periods. In many ad hoc scenarios, a larger warehouse can increase credit consumption, especially when the dominant waste is idle runtime rather than query execution duration.