I’m not sure I’d classify it as a bug. Instances can temporarily go down at any moment for numerous reasons, to account for this instances will keep retrying to connect with an exponential backoff. At what point should an instance assume that another instance is permanently gone?
Perhaps a good start would be adding a status indicator to every community with something like last sync: 1 minute ago.
You can see that an instance/community is gone by visiting the instance directly. In this case at https://lemmy.fmhy.ml/c/imaginarymechas (which obviously won’t work now, as it’s gone).
Whenever you submit a post to a community, first your own instance saves the post locally, then sends it to the instance hosting the community, this instance then sends it to any other instance with users subscribed to the community. When the hosting instance is down, then that step of course fails, resulting in the post being only visible to members of your own instance.
This is just an educated guess, but could it possibly mean that it couldn’t create your post?
This is probably it, the instance they’re communicating with has only 5 instances in their allowlist (look at the bottom of this page).
As that is the instance they’re posting this from, they probably won’t see my comment.
I’m surprised the docker guide still isn’t fixed, the Ansible template got updated to use 2 nginx configs. Take a look at the second one: https://raw.githubusercontent.com/LemmyNet/lemmy-ansible/main/templates/nginx_internal.conf
It splits traffic between the UI and the backend.
It should also be noted that the post will only appear on that kbin instance, and no other instances.