ADA-C01 Question 422
Select 2Manage costs for serverless and AI featuresA Snowflake administrator notices a sudden increase in monthly consumption after several teams begin using serverless features for data pipelines and AI-assisted analysis. Management wants better cost control without blocking legitimate usage. The administrator must reduce the risk of unexpected spend and improve chargeback visibility across teams. Which TWO actions should the administrator take?
- A
Apply resource monitors to the warehouses that invoke serverless features so Snowflake can suspend serverless compute when credit thresholds are reached.
- B
Use tags and cost attribution strategies to associate serverless objects and workloads with business units, then review consumption in ACCOUNT_USAGE views and Snowsight cost monitoring.
- C
Configure budgets and alerts for the relevant serverless and AI-related spend categories so stakeholders are notified before costs become excessive.
- D
Convert all serverless tasks to user-managed virtual warehouses because serverless features cannot be monitored for cost in Snowflake.
- E
Disable query acceleration and AI features at the account level because Snowflake does not provide any native way to observe or govern their costs.
Show answer and explanation
Correct answers: B, C
Explanation
The best answer is to improve visibility and proactive governance for Snowflake-managed spend rather than treating serverless compute like warehouse compute. In Snowflake, resource monitors are designed for virtual warehouses and do not directly control serverless credit consumption. For serverless and AI-related features, administrators should use account usage and billing observability, apply tagging and ownership conventions for chargeback, and configure budgets/alerts to notify stakeholders before spend exceeds expectations. This approach reflects Snowflake best practices for governing Snowflake-managed compute: monitor through ACCOUNT_USAGE and Snowsight, attribute usage with tags and object ownership, and use budgeting/alerting capabilities to reduce the risk of surprise charges.
- A. Incorrect.
Incorrect. Resource monitors apply to virtual warehouses, not to Snowflake-managed serverless compute. A common misconception is that any compute consuming credits can be governed through warehouse resource monitors. However, serverless features such as serverless tasks and other Snowflake-managed compute are billed separately from customer-managed warehouses and are not suspended by warehouse resource monitor thresholds.
- B. Correct.
Correct. Tagging objects and organizing workloads for cost attribution is a practical administrative control for chargeback and accountability. Snowflake provides account-level usage and billing visibility through views in the ACCOUNT_USAGE schema as well as Snowsight cost interfaces. For serverless features, administrators commonly rely on usage views, object metadata, and governance practices such as tagging to understand who owns consumption and where optimization efforts should be directed.
- C. Correct.
Correct. Budgets and alerts are an appropriate way to control the risk of unexpected spend for serverless and AI-related features. Rather than attempting to hard-stop all serverless usage through warehouse-level controls, administrators should configure monitoring and notifications for cost categories and consumers so that teams can respond before monthly charges become excessive. This aligns with cost-governance best practices for Snowflake-managed services.
- D. Incorrect.
Incorrect. Moving everything to user-managed warehouses is not required and is often counterproductive. Serverless features exist to reduce operational overhead and can be cost-effective when used appropriately. The key issue in the scenario is governance and visibility, not the impossibility of monitoring cost. Snowflake does provide ways to observe and analyze serverless consumption.
- E. Incorrect.
Incorrect. This option is too extreme and based on a false premise. Snowflake does provide native observability and governance capabilities for spend, including usage data, cost monitoring in Snowsight, and budget/alerting mechanisms. Disabling useful features account-wide would harm legitimate workloads and does not address the requirement for better control and chargeback visibility.