Attributing queries with QUERY_TAG

Claro's data platform team is investigating an unexpected spike in warehouse credit consumption. WAREHOUSE_METERING_HISTORY shows the surge came from ANALYTICS_WH, but the on-call engineer needs to know which workload or pipeline is responsible. Looking at QUERY_HISTORY, the queries all show the same service account username — there is no further attribution. The team wants a lightweight mechanism to label queries by workload so future investigations can pinpoint the source immediately.

Which Snowflake mechanism lets you attribute queries to a specific workload, pipeline, or application — visible in QUERY_HISTORY?