Read Replicas

Most applications read far more than they write. User dashboards, reporting queries, search results, and API responses all hit the database without changing data. Read replicas let you distribute that load across multiple database instances while your primary handles writes. Ecto makes this simple with its built-in replica support. Define a Repo.Replica module, point it at your replica URL, and call Repo.replica().all() for read operations. Gigalixir provisions replicas with one command and automatically configures replication from your primary.

Unlimited Replicas Per Database: Add as many read replicas as your workload demands. Each replica can be sized independently. Run a small replica for development queries and a large one for analytics. Scale replicas up or down without touching your primary.

Native Ecto Integration: Ecto 3.0+ supports read replicas out of the box. Create a replica repo module, configure it with your READ_REPLICA_URL, and Ecto handles connection pooling separately. Round-robin across multiple replicas with a simple wrapper module.

Independent Scaling and HA: Each replica can have its own size and high availability configuration. Need a beefy replica for reporting but a smaller one for cache warming? Configure them independently. Enable HA on critical replicas without upgrading your primary.

One Command Provisioning: Run gigalixir pg:read_replicas:create and we handle the replication setup. Your replica stays synchronized automatically. No manual configuration of streaming replication, WAL shipping, or failover logic.