Can my humble, single-user instance handle it? Are there any settings I need to be careful about? Should I just backup my server and wing it?
I would suggest you to set Federation mode to seed only. I do not recommend “full” mode for instances under 100 users.
NSFW? Your preference. Auto add? Default is fine.
The option I see is mutual versus seed only. Is the difference that my bot account won’t subscribe to new communities in seed only mode and therefore what I see when viewing All won’t change?
If you seed only, then your bot won’t subscribe. So your All tab will be the same.
Mutual mode can put an additional load on your instance if you are not following the top communities on Lemmy. But I believe it is worth it.
If the extra load becomes a problem, you can disable the bot at any time. With the “Reset subscriptions” button, your bot will unfollow from all communities.
I believe so! It looks like your instance was added and guaranteed on fediseer, and now shows up on Lemmy Federate
My instance doesn’t have many users and we were guaranteed & added as well. I think it’s a pretty painless process for most folks.
The only thing I would keep in mind is changing your site config or domain name could affect federation with other instances
Can my humble, single-user instance handle it?
Maybe?
FWIW this information is based on the experience when using remote hosting not self hosting.
The main issue for us was tuning the database performance. The bot can beat your database to death sometimes and just throwing resources at it won’t necessarily solve it. related low effort meme Our experience was that the bot performance hit wasn’t really noticable immediately but it adds up as more and more communities are added to the database. There also seem to be bursts of database load at times, maybe that’s when new instances join the network? They seem to only last a short time though and haven’t really caused problems.
TLDR- Expect an eventual performance hit. If it’s too much you can always disable the bot or adjust what it does.
did you guys end up caching the db queries to help alleviate it?
I’m not sure it’s been entirely solved. Our host is working through an unrelated hardware problem that hit in the middle of our last round of tuning but things have been behaving better aside from that.
The answer for us has been in adjusting the lemmy pool size to the database and tweaking the available resources to find a happy medium.
