Born out of my other thread, Lack of Prometheus metrics, sometimes, after deploy (host specific?), and given the host-specific issues in both that and Inconsistent/spiky CPU performance even on dedicated CPU? - would it be possible to provide a FLY_HOST environment variable?
This may assist in resolving host-specific issues in future, with Fly users being able to provide a list of specific hosts that are (or had been) working vs. those that aren’t (or have issues - see CPU thread).
This is not currently possible, but we may be able to add it in the future. We can see the hosts your apps are running on, though, so if you just include an app name that’s usually enough!
Thanks for confirming. Though certainly not an immediate requirement I still think it would be valuable for end users in determining potential host-specific issues, in situations where either Fly staff aren’t around to investigate the problem and/or the user isn’t paying for Out of Hours support.
As with the two threads in my opening post - it may have been possible to “fix” the problem(s) by deploying and ending up on another host but with access to a host ID a thread could have been created after the event along the lines of “it worked on x/y/z but not on a/b/c”. This would hopefully assist in tracking down the problem(s).
While looking into my other, possibly host-specific issue, as far as I can tell this currently appears to match the Fly host the app is running on:
echo $FLY_PUBLIC_IP | cut -d ":" -f 4