Restarting machines is the correct way to fix this if your cluster is unhealthy.
Using fly machine commands, you can get a list of machines and then restart them. if you’re on the older platform doing fly restart <your-db-app> should do the trick.
Unfortunately postgres “high availability” has not been highly available, at least not as much as we’d like. It should be able to survive single node failures, but it doesn’t always. That’s why we’ve been working on a new postgres solution.