We’re planning a Google Cloud SQL for PostgreSQL setup with high availability, and I’m unsure what an automatic failover looks like from the application’s point of view. The database endpoint may stay the same, but what happens to existing connections and transactions when the instance switches zones?

If you’re using a connection pool, does it recover cleanly on its own, or do you need to tune retries and connection lifetimes? I’d like to hear what you’ve seen with Cloud SQL connection pools under real failover conditions.