Your application is basically two Fly apps, one for Rails and another for Postgres.
Choose two regions you care about, and pick one as the primary region.
- Launch a high available Postgres cluster on the primary region and adds replica in the secondary region according to High Availability & Global Replication · Fly Docs
- Start at least one machine for your Rails app on each region.
To be clear, considering your comment at Increasing Apps V2 availability - #6 by danwetherald, you don’t need standbys and you won’t get them when deploying your app either.
Just be sure to scale up your Rails app to run in multiple regions.