Nothing should've changed you ability to query other services.

Could you share links to the services you were trying to query so we can take a look at them ? Thanks !

Ponder instances in environments able/baker/charlie in project ecb65c72-931f-428c-8070-5f123a0a60d2

Heya, we actually resolved a small issue with railway->railway communication over the public network earlier today
. could you confirm if you're still running into this issue ?

Yes, I am still having this issue.

Could you collect data from the requests made by your service, and get me an example of what the returned error is ?

Prometheus scrape target shows

Get "https://onlyapes-ponder-able.up.railway.app:80/metrics": http: server gave HTTP response to HTTPS client

this is purely a configuration error, you're making a call to port 80 (http port) to an https service. You should remove the :80 from the request url, or change it to 443 if required

It should be clarified that this has been working without any issue prior to the other day, with this same configuration.

Were you on the new edge proxy prior to this ? This is the only change that might've affected it (old proxy->new proxy).

I'm honestly unsure how https:// + port 80 ever worked

I never explicitly set any proxy settings, I believe I was on the old proxy as I deployed this service several months ago when I don't believe you had rolled out the new proxy.

It's not deliberately set to use those ports/protocols, it's whatever prometheus defaults to. It was working before so your guess is as good as mine haha.

We have been cutting users over to the new proxy so I think you just got hit by the change.

Sorry for any impact !

Explicitly using HTTPS scheme in the Prometheus config seems to have fixed it. Thanks for your help.