A retailer has an expensive query over SUBSCRIPTION_EVENTS with large scans and selective aggregation. Query Acceleration Service is available, but the team wants evidence before enabling it broadly. What should be done? The runbook also names SUBSCRIPTION_EVENTS as the first production object to validate after cutover. The acceptance test includes both current rows and historical correction paths from the same pipeline. The release window is Q2 2025, and the team must choose the Snowflake-first design that minimizes operational risk.
