Mastodon instance not federating correctly

[DidierMalenfant]

I migrated my Mastodon instance from masto.host to a cloudron app and I've noticed that it doesn't seem to be federating correctly.

For example I can only see posts from @macrumors@mastodon.socialon my instance until July 11th and then nothing. If I go directly on mastodon.social all their more recent posts are all there there. I don't see any user facing errors on the website or in the admin section (federation for mastodon.social says "No failures on record").

I added these to my env.production file:

RAILS_LOG_LEVEL=debug
LOG_LEVEL=silly

and rebooted the container. Once the site was back up I searched for @macrumors@mastodon.social in the search bar and grabbed the log after that.

I see this:

2025-08-09T20:45:27Z 2025/08/09 20:45:27 [error] 49#49: *5 connect() failed (111: Connection refused) while connecting to upstream, client: 172.18.0.1, server: , request: "GET /api/v1/announcements HTTP/1.1", upstream: "http://127.0.0.1:3000/api/v1/announcements", host: "malenfant.net", referrer: "https://malenfant.net/deck/search?q=%40macrumors%40mastodon.social"

172.18.0.x seems to definitely be the sub-net that the docker container is on but I don't understand what it's trying to do here.

As a bonus, I'm pretty sure that my posts are not being seen by all my followers either but I only have anecdotal evidence of this at this point.

Anyone with some Masto admin experience able to guide me in the correct direction for diagnosing/fixing this? I'm willing to learn but I don't know where to start...

[nebulon]

Not sure about the federation, but at least in the past, it took quite a bit for new instances to start federating. But I don't have more info on that why this was/is the case.

For the connection error, was this a one-time error? I seem to be able to connect fine on that url and see your mastodon instance. The error as such indicates, that the nginx reverse proxy wasn't able to connect internally to the app, this would happen during app startup, where the container is up, but the app not yet fully functional and thus would be expected at that time.

[DidierMalenfant]

Thanks! as it turns out, in my case it was the cloudflare proxy causing issues. I had disabled it for IPv4 but forgot to do the same for IPv6 so the results/errors were confusing because of this.