Amazon RDS Read Replicas are a feature that allows you to create one or more read-only copies of your Amazon RDS database. These replicas can be used to scale out read operations, improve application performance, and provide additional data redundancy for fault tolerance.
Read Replicas are a strategy to scale horizontally, API calls must be changed in order to use the replicas.
Features
Scalability:
Offload read-heavy workloads to one or more read replicas, reducing the load on the primary database.
Automatic Updates:
Read replicas are automatically updated with changes made to the primary database through asynchronous replication.
Read Replica Promotion:
Promote a read replica to become a standalone database in case of primary database failure or for disaster recovery scenarios.
| Database Engine | Replication Type | Notes |
|---|---|---|
| Amazon RDS for MySQL | Asynchronous | Replication lag may occur; best for read-intensive workloads. |
| Amazon Aurora MySQL | Synchronous for replicas in the same Aurora cluster, Asynchronous for cross-region replicas | Synchronous replication ensures low lag within the cluster. |
| Amazon RDS for PostgreSQL | Asynchronous | Replication lag may occur; useful for read-heavy and analytical queries. |
| Amazon Aurora PostgreSQL | Synchronous for replicas in the same Aurora cluster, Asynchronous for cross-region replicas | Ensures minimal lag within the cluster. |
| Amazon RDS for MariaDB | Asynchronous | Similar to MySQL, replication lag is possible during high write loads. |
| Amazon RDS for Oracle | Asynchronous | Requires Oracle Active Data Guard for replication. |
| Amazon RDS for SQL Server | Not Supported | RDS for SQL Server does not support read replicas. |