We have a query that normally runs for 4 minutes but is currently running for 30 minutes. What would you do?

First, capture the current execution plan using EXPLAIN (or the equivalent in your DBMS) to see how the optimizer is executing the query. Look for full table scans, missing indexes, or expensive joins. Compare the plan to the one from a 4‑minute run to spot differences. Next, check the data volume: has the underlying tables grown or changed distribution statistics? If statistics are stale, run a VACUUM ANALYZE or UPDATE STATISTICS. Verify that indexes are still present and not fragmented. If the query uses a hash join, consider increasing the memory allocation for hash tables or switching to a merge join. Also, examine any recent schema changes or added columns that might affect the query. Finally, test incremental changes: add an index, rewrite the query to use a subquery or CTE, or adjust the parallelism settings. Measure the impact after each change to ensure you’re moving toward the 4‑minute target.