AWS RDS Postgres - Lower instance size cross-region Read Replica for DR site

Creating a cross-region read replica with a smaller instance size for your AWS RDS PostgreSQL database can be a viable disaster recovery (DR) solution, but there are some important considerations to keep in mind:

1. Recovery Time Objective (RTO): Using a smaller instance size for your DR site will likely increase your RTO. When you need to failover to the replica, it may take additional time to scale up the instance to handle your full production workload. This could extend the time it takes to fully recover your services.

2. Performance: A smaller instance size may not be able to handle the same level of traffic or workload as your primary instance. This could lead to performance issues if you need to failover for an extended period.

3. Replication Lag: A smaller instance might struggle to keep up with replication from a larger, more powerful primary instance. This could result in increased replication lag, potentially affecting your Recovery Point Objective (RPO).

4. Cost Savings: Using a smaller instance size can help reduce costs in your DR region, which is a valid consideration for many organizations.

5. Scalability: AWS does allow you to modify the instance class of a read replica. So, in a DR scenario, you could potentially scale up the replica before promoting it to be the new primary. However, this would add to your RTO.

6. Testing: It's crucial to thoroughly test your DR setup, including the process of scaling up the replica and ensuring it can handle your full workload.

While this approach can work and may be suitable for some use cases, particularly where cost savings are a priority, it's not typically recommended as a best practice for critical systems that require minimal downtime and data loss in a disaster scenario.

For mission-critical systems, a more robust DR strategy might involve using a replica of the same or similar size as the primary, or considering other options like Aurora Global Database, which offers faster replication and failover capabilities.

Ultimately, the suitability of this approach depends on your specific requirements for RTO, RPO, and the criticality of the system. It's important to carefully evaluate these factors and test thoroughly to ensure your DR solution meets your needs.

Sources
Creating a read replica in a different AWS Region - Amazon Relational Database Service
How a large financial AWS customer implemented high availability and fast disaster recovery for Amazon Aurora PostgreSQL using Global Database and Amazon RDS Proxy | AWS Database Blog
Community | Cost Conscious Disaster Recovery