COF-C03 Question 199
Single answer2.3 Explain monitoring and cost managementA Snowflake administrator notices that compute costs increased sharply over the last 2 weeks. The company wants to identify which workloads and users are driving the increase without using any external monitoring tools. The administrator needs a Snowflake-native approach that provides historical visibility into warehouse consumption and query activity so the team can take action quickly. Which Snowflake feature should the administrator use first?
- A
Query the ACCOUNT_USAGE views in the SNOWFLAKE database, such as WAREHOUSE_METERING_HISTORY and QUERY_HISTORY
- B
Review the data loading history in COPY_HISTORY because it includes all warehouse credit usage by user
- C
Use Time Travel on the affected database to compare past and current warehouse costs
- D
Inspect the virtual warehouse cache utilization because cache hit rate directly reports credit consumption by role
Show answer and explanation
Correct answer: A
Explanation
To investigate increased Snowflake costs using native capabilities, administrators should start with the SNOWFLAKE.ACCOUNT_USAGE schema, which provides historical metadata for monitoring usage and spend. In this scenario, WAREHOUSE_METERING_HISTORY shows warehouse credit usage over time, while QUERY_HISTORY helps identify which queries, users, and workloads contributed to the increase. This aligns with Snowflake best practices for cost management: first identify where credits are being consumed, then optimize warehouse sizing, auto-suspend settings, query patterns, or user behavior. COPY_HISTORY is too narrow because it focuses on load operations, and Time Travel is unrelated to cost metering. Relevant Snowflake documentation includes the Account Usage views, especially WAREHOUSE_METERING_HISTORY and QUERY_HISTORY, as well as guidance on monitoring credit usage and warehouse activity.
- A. Correct.
Correct. The SNOWFLAKE database includes ACCOUNT_USAGE views that are designed for monitoring and cost analysis. WAREHOUSE_METERING_HISTORY helps identify warehouse credit consumption over time, and QUERY_HISTORY can be used to correlate expensive activity with specific queries, users, and warehouses. This is the most direct Snowflake-native way to investigate a recent cost increase using historical account-level metadata.
- B. Incorrect.
Incorrect. COPY_HISTORY is useful for reviewing data load activity, such as files loaded and load status, but it is not the primary source for overall warehouse credit usage by user or workload. An administrator investigating broad compute cost increases needs warehouse metering and query history, not only load history.
- C. Incorrect.
Incorrect. Time Travel is a data recovery and historical data access feature for tables, schemas, and databases. It does not provide warehouse billing or compute consumption analysis. A common misconception is that any historical investigation in Snowflake should use Time Travel, but cost monitoring relies on metadata views instead.
- D. Incorrect.
Incorrect. Cache behavior can affect query performance and indirectly influence compute efficiency, but Snowflake does not use warehouse cache utilization as the primary source for reporting credit consumption by role. Cost analysis should begin with account usage metadata that explicitly tracks metering and query execution history.