Three discovery tasks when working with PostgreSQL master/replica — and how to solve them

▤ Summary

This article explains how to discover the right PostgreSQL node in a master/replica setup and how to handle failover safely. It compares several approaches, including DNS, libpq multi-host, HAProxy, pgpool-II, Patroni, and a custom pg-status HTTP service. It also shows how to support read-your-writes behavior by tracking WAL LSN values and routing reads to replicas that are sufficiently up to date. The article includes implementation examples in Python and FastAPI and demonstrates how the pg-status service can simplify application-level database routing.

▦ Linked Companies

Docker

$50M to $100M