ARA-C01 Question 421
Select 3Warehouse monitoringA data platform team needs to identify why a shared ETL warehouse has rising credit consumption during business hours. The warehouse is configured as a multi-cluster warehouse in AUTO-SCALE mode with a minimum of 1 cluster and maximum of 4 clusters. Users report that some ETL jobs are queuing intermittently, while finance reports that warehouse costs have increased significantly over the last two weeks. The architect wants to determine whether the increased cost is primarily caused by query concurrency pressure, inefficient queries, or idle time. Which Snowflake sources should the architect use together to make the most accurate assessment?
- A
WAREHOUSE_LOAD_HISTORY to evaluate queued load and concurrency over time, and WAREHOUSE_METERING_HISTORY to compare credit usage patterns
- B
QUERY_HISTORY to identify long-running or poorly performing queries, and WAREHOUSE_METERING_HISTORY to correlate those periods with warehouse credit consumption
- C
LOGIN_HISTORY to determine whether more users signed in during the affected period, and ACCESS_HISTORY to estimate warehouse queuing
- D
SHOW WAREHOUSES to review the current warehouse definition, and DATABASE_STORAGE_USAGE_HISTORY to determine whether storage growth caused warehouse scaling
- E
QUERY_HISTORY together with WAREHOUSE_LOAD_HISTORY to distinguish between concurrency-driven queuing and query-level inefficiency
Show answer and explanation
Correct answers: A, B, E
Explanation
For warehouse monitoring, Snowflake best practice is to correlate warehouse-level telemetry with query-level telemetry rather than relying on administrative or audit views alone. WAREHOUSE_LOAD_HISTORY is used to analyze queued and running load over time and is particularly relevant for diagnosing whether concurrency is causing additional clusters to start in a multi-cluster warehouse. WAREHOUSE_METERING_HISTORY provides the credit consumption needed to validate the cost impact. QUERY_HISTORY provides statement-level detail to identify whether long-running or inefficient SQL is contributing to increased warehouse utilization. In practice, an architect would often use all three together: WAREHOUSE_LOAD_HISTORY for concurrency and queuing, QUERY_HISTORY for inefficient workloads, and WAREHOUSE_METERING_HISTORY for cost correlation. This aligns with Snowflake documentation on Account Usage views for warehouse and query monitoring, including QUERY_HISTORY, WAREHOUSE_LOAD_HISTORY, and WAREHOUSE_METERING_HISTORY.
- A. Correct.
Correct. WAREHOUSE_LOAD_HISTORY is designed for warehouse monitoring and helps assess execution load, queued load, and concurrency behavior over time. This is especially useful for understanding whether the multi-cluster warehouse is scaling because of concurrent demand. WAREHOUSE_METERING_HISTORY shows credit consumption by warehouse, allowing the architect to correlate spikes in load with actual cost increases. Together, these views are a strong choice for identifying whether concurrency pressure is driving additional cluster usage and cost.
- B. Correct.
Correct. QUERY_HISTORY is the primary source for analyzing individual query duration, execution patterns, and potential inefficiencies such as long-running transformations or repeated expensive statements. WAREHOUSE_METERING_HISTORY adds the cost perspective by showing warehouse credit usage during the same time windows. If credit spikes align with inefficient or prolonged queries rather than high concurrency, this combination helps reveal that pattern.
- C. Incorrect.
Incorrect. LOGIN_HISTORY can show authentication activity, and ACCESS_HISTORY is useful for governance and object access auditing, but neither is intended to assess warehouse queuing or cluster scaling behavior. A rise in sign-ins does not reliably explain warehouse load, and ACCESS_HISTORY does not provide warehouse execution load metrics. This option reflects a common misconception that user activity logs can substitute for warehouse performance telemetry.
- D. Incorrect.
Incorrect. SHOW WAREHOUSES returns the warehouse configuration at the current point in time, which may be useful for confirming settings such as AUTO-SCALE and cluster limits, but it does not provide historical monitoring data needed to explain the cost increase. DATABASE_STORAGE_USAGE_HISTORY tracks storage consumption, not compute behavior, and storage growth does not cause a warehouse to scale out. This mixes storage monitoring with warehouse compute monitoring.
- E. Correct.
Correct. QUERY_HISTORY helps determine whether specific statements are inefficient or long-running, while WAREHOUSE_LOAD_HISTORY shows whether the warehouse was under concurrent pressure that caused queued overload or scaling events. Used together, they help separate two common root causes: too many simultaneous queries versus inefficient SQL consuming resources for extended periods.