ARA-C01 Question 392
Single answerScale in/out (multi-cluster warehouse/auto-scaling)A retail company runs hundreds of short BI dashboard queries during business hours and also executes a small number of long-running ELT statements on the same Snowflake virtual warehouse. Users report intermittent queuing between 9 AM and 11 AM, even though the warehouse is already sized as LARGE. The architect wants to reduce query queuing for concurrent dashboard workloads without unnecessarily increasing compute consumption during quiet periods. Which configuration change best addresses this requirement?
- A
Convert the warehouse to a multi-cluster warehouse with MIN_CLUSTER_COUNT = 1, MAX_CLUSTER_COUNT > 1, and auto-scaling enabled
- B
Increase the warehouse size from LARGE to XLARGE and keep it as a single-cluster warehouse
- C
Enable Query Acceleration Service on the warehouse to eliminate concurrency-related queueing
- D
Create a resource monitor with a lower credit quota so the warehouse scales down faster during peak periods
Show answer and explanation
Correct answer: A
Explanation
The key requirement is to reduce queueing caused by concurrency spikes while avoiding a permanent cost increase during off-peak hours. In Snowflake, multi-cluster warehouses are the standard architecture choice for scaling out to support concurrent query workloads. With auto-scaling enabled, Snowflake can start additional clusters when demand increases and shut them down when they are no longer needed. This is more appropriate than simply scaling up to a larger single cluster, which may help some query runtimes but is less targeted for concurrency and keeps higher capacity allocated whenever the warehouse is active. Snowflake documentation and best practices distinguish scale up (larger warehouse size for more compute per query) from scale out (multi-cluster warehouses for more concurrent queries). Resource monitors govern spend, not performance scaling, and Query Acceleration Service is not a substitute for concurrency management.
- A. Correct.
Correct. A multi-cluster warehouse is designed to handle higher concurrency by adding additional clusters when query load increases and reducing clusters when load subsides. Setting MIN_CLUSTER_COUNT = 1 keeps cost lower during quiet periods, while MAX_CLUSTER_COUNT > 1 allows Snowflake to scale out during the morning BI surge. This directly addresses queueing caused by concurrent workloads, especially many short-running queries competing for execution slots.
- B. Incorrect.
Incorrect. Increasing from LARGE to XLARGE scales up a single cluster, which can improve performance for some workloads, but it does not address concurrency as effectively as scaling out with multiple clusters. A larger single-cluster warehouse may still experience queuing when many users submit queries at the same time. It also increases cost continuously while the warehouse is running, including during non-peak periods.
- C. Incorrect.
Incorrect. Query Acceleration Service can help certain eligible queries by offloading portions of query processing, but it is not the primary feature for solving warehouse concurrency queueing. The scenario is specifically about many concurrent BI queries and minimizing cost outside busy periods, which aligns with multi-cluster auto-scaling rather than QAS.
- D. Incorrect.
Incorrect. Resource monitors are used to observe and control credit consumption by issuing notifications or suspending warehouses/accounts when thresholds are reached. They do not control cluster scale-out behavior or reduce queueing. Lowering a quota would more likely interrupt workloads than improve concurrency.