- Nothing stopping the possibility of having redundancy though relays. nostr actually has this by design. If one node goes down, the service might get degraded but never interrupted.
- "self-hosting" does not necessarily mean the act of hosting it yourself. One could, e.g, use a service provider like elest.io, yunohost or my own [0] to run the services. You'd still be responsible for the basic administration, but you'd have actual professional support for uptime, software maintenance, backup, etc.
It feels that mainstream social media has set the expectation that any downtime is big news. That’s, of course, not the case for small social media. People can handle downtime.
If anyone is interested in contributing to small social media, this site is a good guide for the Fediverse side of things: http://runyourown.social/
Wait, what? Aren't we talking about lemmy? The instance owner absolutely does host, in clear, all of its users' content. What am I missing?