COF-C03 Question 202
Single answerCost and warehouse monitoringA Snowflake administrator notices that compute costs increased sharply over the past two weeks. The engineering team says no new warehouses were created, but several BI dashboards now refresh more frequently during business hours. The administrator needs to identify which warehouses are generating the most credits and whether concurrency pressure is contributing to the increase. Which Snowflake source would provide the most direct and relevant information for this investigation?
- A
Query the ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY and ACCOUNT_USAGE.WAREHOUSE_LOAD_HISTORY views
- B
Query the ACCOUNT_USAGE.STORAGE_USAGE and ACCOUNT_USAGE.TABLE_STORAGE_METRICS views
- C
Use SHOW DATABASES and DESCRIBE DATABASE to compare database sizes over time
- D
Query the INFORMATION_SCHEMA.COLUMNS view for all schemas used by the BI dashboards
Show answer and explanation
Correct answer: A
Explanation
For warehouse cost and utilization investigations in Snowflake, the most direct account-level sources are the Account Usage views related to metering and load. ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY is commonly used to analyze compute credits consumed by each warehouse across time periods. ACCOUNT_USAGE.WAREHOUSE_LOAD_HISTORY is used to evaluate warehouse workload characteristics, including whether queries experienced queuing or heavy concurrent demand. In this scenario, more frequent BI dashboard refreshes can increase query volume during business hours, which can raise warehouse credits consumed and potentially create concurrency-related pressure. Reviewing both cost and load together is a Snowflake best practice for determining whether the issue is simply increased demand, a warehouse sizing issue, or a pattern that may benefit from tuning workload scheduling or warehouse configuration. Snowflake documentation on Account Usage views and cost monitoring supports using these views for compute cost analysis and warehouse performance monitoring.
- A. Correct.
Correct. WAREHOUSE_METERING_HISTORY shows credit consumption by warehouse over time, which directly supports identifying which warehouses are driving compute cost. WAREHOUSE_LOAD_HISTORY helps analyze warehouse load, including queued and running load, which is useful for determining whether concurrency pressure or insufficient warehouse sizing is contributing to increased usage. Together, these views are the most relevant sources for investigating rising warehouse costs and performance patterns tied to dashboard refresh activity.
- B. Incorrect.
Incorrect. STORAGE_USAGE and TABLE_STORAGE_METRICS are focused on storage consumption and storage-related costs, not warehouse compute usage. A candidate might choose this option because total Snowflake cost includes storage, but the scenario specifically asks about increased compute costs and warehouse activity.
- C. Incorrect.
Incorrect. SHOW DATABASES and DESCRIBE DATABASE can help inspect database metadata, but they do not provide warehouse credit usage or warehouse load/concurrency details. This option reflects a common misconception that rising overall platform cost must be tied to database growth rather than compute activity.
- D. Incorrect.
Incorrect. INFORMATION_SCHEMA.COLUMNS contains metadata about table columns and is unrelated to warehouse credit consumption or warehouse load. Someone might pick this if they are thinking about tracing BI dashboard dependencies, but it does not directly answer which warehouses are consuming credits or whether concurrency is affecting warehouse behavior.