You currently run a Cloud SQL instance in a single zone and need high availability. What should you do?

To achieve high availability for a Cloud SQL instance, you need a failover replica. This replica automatically takes over if the primary instance becomes unavailable. Placing the failover replica in a different zone within the same region ensures resilience against zonal outages without introducing significant latency from cross-region replication. Read replicas are for scaling read operations and do not provide high availability for the primary instance. Creating replicas in a different region would introduce higher latency for synchronous replication and failover, which is generally not preferred for high availability within a single application deployment unless it's for disaster recovery.