postgres: Don't infinitely keep restarting when DB couldn't start

[srid]

This just goes on an infinite loop trying to restart when it will never succeed. Should we not crash the whole process?

cf.

services-flake/nix/services/postgres/default.nix

Line 337 in 113f8cb

availability.restart = "on_failure";

[srid]

This can be reproduced by simply running two instances of nix run ....

As an aside, perhaps all services-flake service should check for lockfile on its data directory and crash immediately if it was acquired.