Add timeouts to my database queries. Find where queries run and check whether any statement or execution timeout is set; most setups have none, which means a slow query runs forever. Configure: a statement timeout at the database or driver level (something like 5-10 seconds for interactive web queries, longer only for known slow jobs like exports); error handling for timeout errors that returns a clean failure to the user instead of a hung request; and logging that records which query timed out so slow queries get found instead of silently retried by users refreshing. Separate interactive queries from background jobs: exports, reports, and migrations legitimately need longer, so give them their own timeout instead of loosening the global one. Don't change query logic in this pass. Deliver: the timeout configuration, per-context timeout values with reasoning, the error handling for timeouts, and verification: run an intentionally slow query and confirm it gets killed at the timeout instead of hanging.