When querying for “legal” and “law” on lemmyverse.net, it’s starkly clear that many communities are missing from the database. I get far more results when querying on specific instances. So what’s the problem? Is it no longer crawling?
The crawler has been wonky for a week or more. If you check the recent posts at !trendingcommunities@feddit.nl you’ll see the list has been very unpredictable lately as some of the bigger instances aren’t being reported by the bot. Simultaneously there seems to be a federating bug that no one has solved yet. It was/is affecting .world communities federating to lemmy.ml and it’s been causing communities on walledgarden and some other instances to fall in and out of sync with each other.
The law communities I was expecting to find are old and on small instances. When I searched on just the domain portion of a small instance, Lemmyverse finds only like 10% of the communities on that node.
You would expect LV to not forget a community once it is in the DB. But it must also have a removal mechanism for communities that are deleted. It’s as if the removal detection is getting false positives for removals.
I understand the confusion.
Maybe I should’ve been clearer that lemmyverse performance has been unreliable, not just the crawler. Entire instances, communities, and mixes of the two have disappeared and reappeared between scans. It also hasn’t been reliably picking up changes to instances it already knows about.
TLDR lemmyverse has something wrong with it.
