← case study
SQL Performance Lab
Five scenarios, each asked twice. Both versions answer the same question and return the same rows; only the way they are written changes. Press run and both execute against a real PostgreSQL database at that moment — the timings and the execution plans below are measurements, not fixtures.
The queries are fixed on the server; the only thing your browser sends is which scenario to run. The database user is read-only, every statement has a five-second budget, and runs are rate limited — it is a public endpoint executing SQL, and that is what it takes to leave one open.
waking the service and the database — up to a few seconds. Both scale to zero between visits to keep this demo free.