500 error, no logs about phoenix

[prologic]

Hey! Thanks for replying! This sadly (shame on me) was a red herring on my part. I'll share what happened though, as it was this Phoenix instance where I first discovered this.

After hours of debugging, I eventually landed on a random Traefik issue traefik/traefik#3163 that happened to mention turning up the logging level to DEBUG. I did that and discovered the following error(s) that were only visible in debug logs :/

2024-03-23T05:35:42Z" level=debug msg="Error calling http://authelia:9091/api/verify?rd=https://auth.mydomain.tld. Cause: Get \"http://authelia:9091/api/verify?rd=https://auth.mydomain.tld\": dial tcp: lookup authelia on 127.0.0.11:53: no such host" middlewareName=authelia@docker middlewareType=ForwardedAuthType

I have a few service (such as Phoenix) behind Authelia's Proxy auth and using Traefik's Docker-based middleware. Unfortunately the middleware has disappeared "somehow" despite the service still running just fine. This was resulting in 500 erro(s) with zero content, requests were not reaching any backend with this middleware configured.

So adding additional logging won't really help in this cse, but I did notice that we don't really log errors. So my fork addes some additional calls to logrus.WithError(err).Error() in a few places.

[prologic]

No amount of restart, or rebuild or redeployment has worked so far. It's really weird.

[ordinary-dev]

Hey!

Thanks for the bug report, I'll add more logs. The project is unpolished in many aspects, but I use it myself and have not noticed such problems.

Judging by the fact that a 500 error is at least returned, the server starts successfully. I'll check the request handlers.