We are launching support for Postgres Read Replicas, a feature that allows you to distribute load across databases, serve data closer to users, provide data redundancy, and run complex queries without affecting your primary database. You can create read replicas in any of our 12 supported regions, with each project supporting up to two replicas. Read replicas continuously replicate data from a primary database, but replication is asynchronous and may result in some delay. We are working on making it easier to launch and manage read replicas, including automatic failover and geo-based load balancing. The feature will be priced like regular databases for compute and storage, with additional costs for WAL transfer between the primary and read replica. Sign-ups for early access of self-serve Read Replica support are now open, with progressive rollout to all paid plans over the next few months.