Execute an Aurora blue/green deployment for a zero-downtime schema or version upgrade

Steps

1. From the RDS Console select the Aurora cluster, choose Actions → Create Blue/Green deployment; Aurora provisions a green cluster and sets up logical replication from blue to green
2. Apply schema changes, engine version upgrades, or parameter group changes on the green cluster while the blue cluster continues serving traffic
3. Monitor replication lag between blue and green in the console or via CloudWatch; wait for lag to near zero before switching
4. Initiate switchover: Actions → Switch over; Aurora validates health checks, waits for replication to catch up, then atomically renames blue and green endpoints (the switch typically completes in under a minute)
5. Validate the application against the new primary (formerly green); if issues are found, the old blue cluster is still available until you explicitly delete it
6. Delete the blue cluster once validated to avoid incurring double storage and instance costs