Slow query pressure
Ranks pg_stat_statements by total time, mean time, temp writes, and disk reads.
Generate a practical Postgres performance report from local collector data: slow queries, missing indexes, unused indexes, table bloat, autovacuum drift, and expensive sorts.
npx postgresaudit collect --url "$DATABASE_URL" --out audit.jsonBootstrap the first admin account, then sign in to use the secure audit workbench.
Sign in before uploading collector output. Reports may contain private schema and query data.
The collector JSON can expose schema names, slow queries, and table sizes. Login keeps the audit route private while the landing page remains indexable.
This runs once. The first user becomes admin and bootstrap closes after that.
Ranks pg_stat_statements by total time, mean time, temp writes, and disk reads.
Surfaces scan-heavy tables, low-usage indexes, and likely composite-index candidates.
Checks row estimates, dead tuple ratios, autovacuum timestamps, and relation sizes.
Use a restricted PostgreSQL user and export audit.json locally.
Paste or upload the collector output into the secure report workbench.
Get prioritized findings with evidence, recommendations, and risk notes.
Yes. The collector runs metadata and statistics queries, sets read-only session mode, and never executes optimization SQL.
It works without pg_stat_statements, but query-level findings are much better when the extension is enabled.
No. The MVP only collects signals and generates a report. Index creation, VACUUM changes, and setting changes remain manual review steps.