QueryLint

QueryLint is a web app (with optional CLI) that continuously analyzes your cloud data warehouse query history and flags the small set of patterns that reliably drive surprise costs and slowdowns: unbounded scans, missing partition filters, accidental cross joins, inefficient UDFs, and repeated ad-hoc exploration that should be materialized. It connects read-only to Snowflake, BigQuery, and Redshift, builds a baseline of “normal” spend and performance per team, and then generates actionable fixes: rewritten SQL suggestions, recommended clustering/partitioning keys, and candidate materialized views. It also produces lightweight “guardrails” you can enforce via PR comments (dbt/SQL repo) and scheduled alerts (Slack/email) when a query is likely to exceed a cost or runtime threshold. This is not a generic observability dashboard; it’s a focused linting and prevention tool aimed at reducing waste quickly.