high-availability

Surviving a Patroni failover with logical replication

Your Patroni cluster does exactly what you built it to do. A node dies at 3 a.m., a new primary is elected in a few seconds, the application reconnects, and nobody gets paged. Your HA setup works. And(…)