Time out when connecting to app instances via wireguard tunnels

Hi there, when I try to connect to other app instances from one instance (via DNS lookup of global.APP_NAME.internal and then connect to one of the returned IP addresses) it timeouts. I can reproduce this from the command line via fly ssh console.

Is this a known issue or something I might be doing incorrectly?

Thanks!

Is the IP you’re connecting to still alive? Try running fly ips private and see if it exists in the output.

There is a bug that keeps old IPs in DNS in some cases.

Hey Kurt, yep that’s exactly what was happening…except it looks like all the IPs were nonexistent and unreachable. I cannot reproduce this now though (it looks like some of the IPs are still old, but the new ones are now showing up too). I found this issue because I forgot to set a timeout when making a request to another instance. As long as one of the requests succeeded the program would continue, but if all timeout it would stall. My assumption now is that if I cannot reach an instance in <5s it’s probably unreachable…and if no instances are reachable the instance will assume it’s alone in the world :slight_smile: