The problem is most likely that Ask Burggit was not discovered by sh.itjust.works at that time. Someone has to search for it on sh.itjust.works, and then wait a while. People say you have to search the full URL instead the the !community@burggit.moe style search, for sh.itjust.works specifically for some reason.
For whatever reason, instances which are federated to each other to not send lists of communities. Someone has to manually search for each foreign community individually before it is discovered by the instance, and then there is a process of (identifying? copying? archiving?) all the posts in that community which can take 30+ minutes before it’s visible. That’s all my second-hand understanding of the process anyway.
Yeah, and then we were hit by a bunch of rate limiting errors because we got hit by a bunch of federation requests because we got discovered, and I refused to proxy real IPs to our backend for privacy reasons until now, rofl.
This is also my understanding of it, part of it makes sense due to like, resource usage reasons. If your members don’t have an interest in a community, there’s no need constantly getting fetch data for all available communities and posts/comments in them. I do feel like there’s probably a better way of doing it, though I can’t think of one at the moment, so maybe there isn’t. Lemmy is still early development software, so hopefully it will continue to improve in this and other ways as things continue to move forward.
A better way would be on demand API requests. API is there. Very little cost in snatching that when a user goes into a thread and caching it if needed.
The problem is most likely that Ask Burggit was not discovered by sh.itjust.works at that time. Someone has to search for it on sh.itjust.works, and then wait a while. People say you have to search the full URL instead the the !community@burggit.moe style search, for sh.itjust.works specifically for some reason.
For whatever reason, instances which are federated to each other to not send lists of communities. Someone has to manually search for each foreign community individually before it is discovered by the instance, and then there is a process of (identifying? copying? archiving?) all the posts in that community which can take 30+ minutes before it’s visible. That’s all my second-hand understanding of the process anyway.
Yeah, and then we were hit by a bunch of rate limiting errors because we got hit by a bunch of federation requests because we got discovered, and I refused to proxy real IPs to our backend for privacy reasons until now, rofl.
This is also my understanding of it, part of it makes sense due to like, resource usage reasons. If your members don’t have an interest in a community, there’s no need constantly getting fetch data for all available communities and posts/comments in them. I do feel like there’s probably a better way of doing it, though I can’t think of one at the moment, so maybe there isn’t. Lemmy is still early development software, so hopefully it will continue to improve in this and other ways as things continue to move forward.
A better way would be on demand API requests. API is there. Very little cost in snatching that when a user goes into a thread and caching it if needed.