Microservices aren't a silver bullet. There's likely quite a lot that can be done until we need to split some parts out, and once that happens I expect that federation would be the thing to split out as that's one of the more "active" parts of the app compared to logins and whatnot.
The instance I'm replying from is a 5 eur/mo box from Hetzner.
Your main concerns are gonna be active user count & storage space. Especially if you decide to allow image or god forbid video uploads. Having a bunch of inactive users aren't going to affect costs that much as long as they don't have, like, a milion subscriptions. (If they're all subscribed to the same community things will "deduplicate")
It's generally more like "Steve's 10 eur/mo cloud server in which they run ten other things next to Lemmy, which is written by two devs and barely held together by duct tape and prayers"
But that doesn't change the overall point.
With how unreliable tallying votes over federation is, we're kinda get vote fuzzing "for free" right now.
Any admin worth their salt's gonna defederate them and proudly wear the Misfit Loser Zealots label[^1]. The only people who'll federate with them are the naive techbros and those who only care about how much users they have, compared to, idk, being committed to creating a good community.
https://fedipact.online is already gaining steam with the Mastodon side of the fediverse.
[^1]: Seriously the markdown guy couldn't've picked a better description if he tried.
On the software side, we already have PeerTube. It's just the logistics of hosting video are way too expensive for most people to be able to cover:
- You need drives to store all those videos, preferably in several quality and codec variants so everyone can watch them.
- You need the bandwidth to serve all those videos. PeerTube can "smooth over" the initial new upload bump by using WebTorrents, which is the least worst solution if you quietly ignore all the "but muh IP address" people, but once people stop watching at the same time, you're back to square one.
- Transcoding requires powerful and specialized hardware. Nobody in their right mind will serve videos the same way they're uploaded, especially with the rise of new codecs like VP9 or god forbid AV1, which you simply can't encode on a consumer CPU unless you're fine waiting hours/days for a single upload to go through.
Feel free to re-crop and adjust as required.
EDIT: This is getting dangerously upvoted so here is the Krita source file for you to fuck around with should you need it: https://it.w.on-t.work/uJ1i9p.zip
Most of Beehaw's blocks are "generic ActivityPub assholes", which, before the Reddit migration, was really just the worst of the worst of Mastodon, Pleroma, Soapbox, and Miss/Calc/???Key instances, with the occasional PeerTube thrown in.
They likely just imported one of the common blocklists and moved on with their lives, which really should be "how to secure your community 101" but most Lemmy admins haven't seem to have gotten the memo yet.
I'm patiently waiting for the day those assholes realize most of Lemmy is open ground for them to shit in, boy that's gonna be a fun few days.
If only ISPs in my country decided to actually give a shit about IPv6 instead of deciding to NAT even harder.
At this point aren't all that excess processing losing them money? Like, you can still sell IPv4 if that's how you make a profit but what reason do you have to not just click the "turn on IPv6" button that is probably there if your networking equipment is made in the last century or so????
it's it even possible to read this in any voice other than dankpods's
respect to our hard workers, tirelessly working the content mines, all for nothing but a handful of up votes