PostgreSQL Performance Toolkit
CompletedA collection of SQL scripts, pgBadger dashboards, and Python automation tools for diagnosing and resolving slow-query issues in high-traffic PostgreSQL clusters.
Tech Stack
Built over several years of on-call incident response, this toolkit consolidates the most impactful diagnostics for PostgreSQL performance engineering. It targets clusters running ≥500 TPS where off-the-shelf monitoring misses application-specific patterns.
The toolkit includes: a slow-query analyser that correlates pg_stat_statements data with EXPLAIN ANALYZE output and surfaces index candidates; a bloat estimator for tables and indexes with automated VACUUM scheduling recommendations; and a Python CLI that generates weekly performance reports in Markdown, ready for engineering review.
Used in production environments handling up to 2 TB databases. The index recommendation module alone reduced average query latency by 40 % on the most contended tables during peak hours.
Other Projects
Internal data cataloguing and lineage platform integrating dbt, Metabase, and Supabase to give a 200-person org a single source of truth for 300+ data assets.
Serverless appointment-slot monitor for Australian visa applications — sends real-time alerts when new slots open, built on Playwright, Vercel Cron, and Supabase.
Enterprise-grade bidirectional data pipeline connecting NetSuite ERP and Boomi AtomSphere, processing 50 k+ transactions per day with sub-3 s latency.