ADA-C01 Question 388
Single answerMonitor accounts and usageA Snowflake administrator needs to identify which users and warehouses are driving a sudden increase in compute spend over the last 14 days. The company wants an analysis that can be refreshed daily and queried with SQL inside Snowflake, without exporting data to another tool. Which approach should the administrator take?
- A
Query the SNOWFLAKE.ACCOUNT_USAGE.QUERY_HISTORY and WAREHOUSE_METERING_HISTORY views, then correlate query activity, warehouse usage, and users over the desired time window
- B
Use only SHOW WAREHOUSES and SHOW USERS commands because they return the most current data and include historical credit consumption by user
- C
Query INFORMATION_SCHEMA tables in each database because they provide the authoritative account-wide billing history with the longest retention
- D
Review the COPY_HISTORY view to determine which users and warehouses consumed the most compute credits across all workloads
Show answer and explanation
Correct answer: A
Explanation
For monitoring account usage and investigating increased spend, Snowflake best practice is to use the SNOWFLAKE.ACCOUNT_USAGE views. In this scenario, the administrator needs historical, SQL-queryable data inside Snowflake that can be refreshed regularly. QUERY_HISTORY helps identify which users ran expensive or high-volume workloads and which warehouses they used, while WAREHOUSE_METERING_HISTORY shows credit consumption by warehouse over time. Together, these views support root-cause analysis of compute spikes. SHOW commands are primarily for current metadata inspection, not historical spend attribution. INFORMATION_SCHEMA is useful in many administrative workflows, but for account-level historical usage and billing-style analysis, ACCOUNT_USAGE is the more appropriate source. Snowflake documentation for Account Usage and Organization/Account monitoring consistently positions ACCOUNT_USAGE as the core source for historical monitoring and usage analysis.
- A. Correct.
Correct. The SNOWFLAKE.ACCOUNT_USAGE schema is designed for account monitoring and historical analysis. QUERY_HISTORY provides query-level details such as user, warehouse, execution timing, and other metadata, while WAREHOUSE_METERING_HISTORY provides warehouse credit consumption over time. Joining or correlating these views is a practical way to investigate which users and warehouses contributed to increased spend during a defined period. This approach also satisfies the requirement to keep the analysis in Snowflake and refresh it regularly with SQL.
- B. Incorrect.
Incorrect. SHOW commands are useful for current object metadata and administrative inspection, but they are not the right source for historical spend analysis by user. SHOW WAREHOUSES does not provide user-level historical credit consumption, and SHOW USERS does not contain billing history. A common misconception is that SHOW output is sufficient for monitoring trends, but it is not intended to replace ACCOUNT_USAGE views for historical usage analysis.
- C. Incorrect.
Incorrect. INFORMATION_SCHEMA contains useful metadata and some usage-related table functions, but it is not the primary source for long-term, account-wide billing analysis. ACCOUNT_USAGE is specifically intended for historical monitoring across the account, with views such as QUERY_HISTORY and WAREHOUSE_METERING_HISTORY. Choosing INFORMATION_SCHEMA here reflects the misconception that database-scoped metadata is equivalent to account-level usage monitoring.
- D. Incorrect.
Incorrect. COPY_HISTORY is limited to data-loading activity and does not represent all compute-consuming workloads. It would miss SELECT, INSERT, UPDATE, transformation, task-driven, and other query activity that may be responsible for the spend increase. This option is plausible because loading can consume credits, but it is too narrow for identifying overall compute drivers.