COF-C03 Question 56
Single answer1.4 Configure virtual warehousesA retail analytics team runs thousands of short BI queries during business hours. Users complain that dashboards become slow whenever several analysts refresh at the same time. The Snowflake administrator wants to improve concurrency for this workload without manually resizing warehouses throughout the day, while also avoiding unnecessary compute cost overnight. Which configuration change best addresses this requirement?
- A
Configure the virtual warehouse as a multi-cluster warehouse with auto-scale enabled and auto-suspend set to a short interval
- B
Increase the warehouse size to 6XL and disable auto-suspend so queries do not queue during peak periods
- C
Enable Query Acceleration Service on the existing warehouse and keep it running 24/7
- D
Convert the warehouse to a Snowpark-optimized warehouse so concurrent dashboard queries have more memory available
Show answer and explanation
Correct answer: A
Explanation
For high-concurrency BI workloads, Snowflake best practice is to use a multi-cluster warehouse rather than simply increasing the size of a single cluster. Multi-cluster warehouses are specifically intended to support many concurrent users and queries by automatically adding clusters when needed. Auto-scale helps match compute to demand, and auto-suspend helps minimize cost when the warehouse is idle. In contrast, larger single-cluster warehouses are more useful for improving throughput of larger queries, not handling bursts of concurrent small queries. Snowflake documentation on virtual warehouses and multi-cluster warehouses emphasizes using multi-cluster configuration for concurrency-sensitive workloads such as dashboards and reporting.
- A. Correct.
Correct. A multi-cluster warehouse is designed to handle concurrency by adding clusters when many queries are submitted at the same time. Auto-scale allows Snowflake to start additional clusters as needed and reduce them when demand drops, which fits a BI workload with bursty daytime usage. Setting auto-suspend to a short interval also helps control costs during idle periods, especially overnight.
- B. Incorrect.
Incorrect. Increasing to a much larger single-cluster warehouse may improve performance for some resource-heavy queries, but it does not address concurrency as effectively as a multi-cluster warehouse. Disabling auto-suspend would also increase cost because the warehouse would continue running when no queries are being executed.
- C. Incorrect.
Incorrect. Query Acceleration Service can help accelerate parts of eligible queries, but it is not the primary feature for solving warehouse concurrency issues caused by many simultaneous dashboard requests. Keeping the warehouse running continuously also conflicts with the requirement to avoid unnecessary overnight cost.
- D. Incorrect.
Incorrect. Snowpark-optimized warehouses are intended for workloads that need more memory and compute resources for Snowpark jobs, such as machine learning or data engineering tasks. They are not the best choice for improving concurrency for many short BI queries.