New issue: postgres instance not reachable anymore from app :nxdomain

Hi, my postgres instance is not anymore reachable through the .internal domain, when my app tries to connect, it says :nxdomain

2022-06-15T13:36:48Z app[**] fra [info]13:36:48.269 [error] Postgrex.Protocol (#PID<0.1967.0>) failed to connect: ** (DBConnection.ConnectionError) tcp connect (***.internal:5432): non-existing domain - :nxdomain

Same now goes for both organisations that I belong to – both of my apps are down (since a couple of minutes the second followed)

Emergency maintenance had to be performed on 2 hosts in FRA and something went wrong when instances were coming back up.

Status page: Status - FRA Emergency Maintenance

1 Like

Ah, I see now – thank you! :slight_smile:

They are now coming back up, slowly.

1 Like

I believe the situation has been resolved for your app as of ~10 minutes ago.

1 Like

One note here: if you’re running databases with a single node, and this kind of outage is a pain in the ass, you should think about adding a second node. Our infrastructure is designed to be reliable for sets of 2+ VMs. It’s less reliable than, say, DigitalOcean for a single VM + disk.


@kurt thank you for the note, that’s helpful! In the moment uptime is not crutial since we’re only testing but I will keep that in mind for when we‘re live. cheers

By the way, one of my apps still did not have connectivity to postgres with :nxdomain – I managed to make it work through flyctl restart [app name]

You restarted the app connecting to postgres or the postgres app itself?

@jerome I restarted the postgres app itself