Zero-Downtime PostgreSQL Major Version Upgrades in Containers: The Problem Nobody Talks About

Minor PostgreSQL upgrades are straightforward in containers, but major version jumps are harder because pg_upgrade needs both old and new binaries available on the same host.

Traditional paths like dump-and-restore or logical replication can work, but they are slow, operationally complex, and often mean longer downtime.

The post introduces pg-upgrade, a Docker-native workflow that packages the migration steps into a reproducible upgrade process.

It is positioned as a practical way for self-managed PostgreSQL teams to reduce risk and validate upgrades in CI before production.