Physical replication copies the whole cluster, byte for byte. Logical replication works at the row level instead: you pick the tables, the changes stream to any number of subscribers, and the subscriber can be a different PostgreSQL version, a different schema layout, or a different system entirely. That flexibility is why it sits underneath selective replication, data integration, and near-zero-downtime upgrades. It is also why it has more moving parts than primary_conninfo.
This guide covers logical replication from the first publication to production operation: the publish-subscribe model and its components, then monitoring, data copy strategies, decoding internals, and security. Four more chapters are planned, moving from conflict resolution and performance through major version upgrades, blue/green schema migrations, and change data capture.
The guide is a work in progress. If you want a note when the next chapter lands, the newsletter is the channel:
