ADA-C01 Question 361
Single answerSet up a dashboard to monitor Snowflake costsA Snowflake administrator has been asked to build a cost-monitoring dashboard that shows daily spend trends by warehouse and helps identify which workloads are driving compute charges. The company wants the dashboard to be based on Snowflake-native metadata, with SQL queries that can be refreshed daily. Which approach should the administrator use?
- A
Query ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY and join it with ACCOUNT_USAGE.QUERY_HISTORY or TAG references as needed to attribute compute consumption by warehouse and workload over time.
- B
Query INFORMATION_SCHEMA.QUERY_HISTORY for the account and sum execution time by user, because execution time directly represents Snowflake credit consumption.
- C
Use ACCOUNT_USAGE.STORAGE_USAGE exclusively, because all Snowflake charges, including warehouse compute, are consolidated there for dashboarding.
- D
Query ACCOUNT_USAGE.LOGIN_HISTORY and aggregate session duration by role to estimate daily warehouse spend by workload.
Show answer and explanation
Correct answer: A
Explanation
To monitor Snowflake costs with a dashboard, administrators should rely on Snowflake metadata views that directly represent billed usage. For warehouse compute, ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY is the key source because it provides warehouse credit usage over time. To make the dashboard actionable, this can be combined with ACCOUNT_USAGE.QUERY_HISTORY and organizational tagging strategies to analyze which warehouses, users, roles, or workloads are associated with spend patterns. This aligns with Snowflake best practices for cost monitoring and chargeback/showback reporting. By contrast, query runtime, login duration, or storage-only views are indirect or incomplete proxies and should not be used as the authoritative source for compute cost dashboards.
- A. Correct.
Correct. ACCOUNT_USAGE.WAREHOUSE_METERING_HISTORY is the primary Snowflake metadata view for warehouse credit consumption over time and is appropriate for building a dashboard of daily compute spend trends. To understand which workloads are contributing to that spend, administrators commonly correlate warehouse usage with ACCOUNT_USAGE.QUERY_HISTORY, warehouse names, users, roles, or business tags applied to warehouses or related objects. This is a practical, Snowflake-native method for cost observability.
- B. Incorrect.
Incorrect. INFORMATION_SCHEMA.QUERY_HISTORY is useful for recent query-level operational analysis, but query execution time is not the same as billed credits. Warehouse billing depends on warehouse runtime and size, including periods where the warehouse is running without actively executing a query. Summing query duration can therefore significantly misrepresent actual compute cost.
- C. Incorrect.
Incorrect. ACCOUNT_USAGE.STORAGE_USAGE is for storage-related usage and charges, not warehouse compute consumption. A dashboard based only on STORAGE_USAGE would omit the core requirement of showing spend trends by warehouse and identifying compute-heavy workloads.
- D. Incorrect.
Incorrect. LOGIN_HISTORY tracks authentication events, not warehouse metering or credit billing. Session duration by role does not map reliably to warehouse credit usage because users can log in without running queries, and multiple users can share warehouses with varying levels of compute consumption.