Reduce the size of our RDS instances

### Context

https://github.com/AlexsLemonade/refinebio/issues/2042

Idea 4 and 5

### Problem or idea

Reduce the size of the staging RDS instance. It currently is just as big as the prod one and it really doesn't need to be. We could probably use an instance a quarter the size and still be fine.

### Solution or next step

We should be careful about how we do this so that we don't kneecap ourselves when we do want to larger-scale processing again. I think that before this goes to prod that we should run in staging or dev at the largest scale we want to be capable of and monitor the database performance to make sure we're still okay with a smaller instance.

In combination with https://github.com/AlexsLemonade/refinebio/issues/2171, I think our new cap might be at ~5TB of RAM, because I think with our currrent sizes we've gone as big as 20TB of RAM.

Start with the linked issue 2042 and locate the staging RDS configuration; also review issue 2171 for the related capacity change. Run the largest-scale workload the project needs in staging or dev while monitoring database performance. Done means a smaller staging instance supports that workload without unacceptable performance and can be safely considered for production.

Written by the indexing model from the issue text.