Here's an idea:
- An RDS proxy in the EKS cluster that can send read requests to the replicas, and write requests to the master (which will mean less load on the RDS writer, and therefore we can use smaller clusters and save $)
- The proxy is also AZ-aware, and can talk to the replica that's in its own AZ. This will save on data transfer costs
Does that sound like something that might be added to the roadmap at some point?
Here's an idea:
Does that sound like something that might be added to the roadmap at some point?