QueryTriage

QueryTriage is a web app (with optional desktop agent) that pinpoints why dashboards and ad-hoc BI queries are slow, expensive, or timing out across Snowflake/BigQuery/Databricks. It ingests query history, warehouse/resource metrics, and BI tool metadata, then groups recurring “bad patterns” (runaway joins, missing clustering/partitioning, over-wide selects, chatty dashboards, inefficient semantic layer SQL). It produces a prioritized backlog with estimated cost savings, suggested rewrites, and safe rollout steps. The AI layer explains issues in plain language, proposes alternative SQL, and generates dbt/warehouse tuning recommendations, but keeps humans in control with diffs and approval gates. Realistically, this won’t replace a data engineer; it reduces firefighting time and makes performance work measurable. Best fit is teams already paying meaningful warehouse bills and suffering stakeholder pressure from slow dashboards.