[SOLVED] For Disaster Recover, how much faster is it in practice to failover an Aurora Global database vs promote an RDS read replica?

For Disaster Recover, how much faster is it in practice to failover an Aurora Global database vs promote an RDS read replica?

From reading the docs, it's not clear if Aurora Global Database has a lower RTO vs promoting an RDS read replica. Does anyone know if there is a difference?

AWS docs say: "For an Aurora Global Database, RTO can be in the order of minutes."

For promoting RDS read replica, RDS docs say "The promotion process takes a few minutes to complete. When you promote a read replica, replication is stopped and the read replica is rebooted. When the reboot is complete, the read replica is available as a new DB instance."

Solution

Both promoting read replica (with in the same region) and Aurora Global Databases failover takes few minutes, both have different use cases/purpose.

Global database provides regional level failover protection (Example: Failover from US to UK) compare to read replica that you referred is AZ level (which are with in 60 mile range)

Also, Global database provides low-latency local reads in different Region