By looking at docs, I saw that the Fly team did a fantastic job with a turn-key high availability for PostgreSQL.
Please correct me if I’m wrong, but it’s completely automatic for a project with 2+ instances.
My question here is whether there are similar conveniences for PITR?
Maybe I’m asking for too much. In that case, perhaps someone here already implemented PITR for PostgreSQL and is willing to share, if not a recipe, but maybe even a list of the most intricate gotchas and road bumps that you run into.