Migrate Amazon Aurora PostgreSQL across major versions with active Debezium CDC connectors using native logical replication

This article demonstrates how to migrate Amazon Aurora PostgreSQL across major versions while maintaining active Debezium CDC connectors using native logical replication to avoid full re-snapshots.

• Use PostgreSQL native logical replication to bridge source and target clusters during major version upgrades
• Execute a brief cutover sequence: disable subscription, synchronize sequences, create Debezium slot on target, redirect connector
• Keep the same topic.prefix so Kafka consumers continue reading from unchanged topics with minimal pause
• Set snapshot.mode=never to skip initial snapshot and resume from current LSN without backlog
• Monitor flush_lag and WAL growth post-cutover; connector can report RUNNING while silently accumulating WAL
• Handle failure scenarios: Aurora failover, connector stalls, large messages, DDL changes, and long-running transactions
• Database-side cutover completed in ~2 seconds in testing; total consumer pause depends on connector startup time
• Requires low-latency connectivity between clusters; same VPC/AZ recommended, cross-AZ viable, cross-Region not recommended

This pattern enables zero-downtime major version upgrades for Aurora PostgreSQL with active CDC pipelines by using logical replication as a bridge, eliminating the need for full Debezium connector re-snapshots that can take hours per database.