Google Associate Cloud Engineer: Cloud SQL High Availability

Cloud SQL high availability is easiest to understand as a regional failure-control mechanism rather than as a generic promise that a database can never go down. A high-availability Cloud SQL instance places a primary and a standby in different zones of one region and synchronously replicates writes to persistent storage in both zones before reporting the transaction as committed.

If the primary or its zone becomes unavailable, Cloud SQL can fail over to the standby and continue serving through the same instance identity and IP address after clients reconnect. That model removes a large amount of database infrastructure work, but it does not remove architecture work.