Considering how much bullshit I have to deal with regarding Pleroma maintenance on Schwartzwelt's servers I'm thinking of finally figuring out containerization for Pleroma. If nothing else I hope it can prevent the system from crashing and burning for indiscernible reasons.
@Tadano@silverpill seconding this, rendering without being saved to server disk is exactly what I was thinking. Negates the risk of having cheese pizza or other such garbage being federated to disk while saving on space. A whitelist for instances to cache from for better speed would be super cool too! Though I don't know how complex that'll be to implement.
Apologies for the lack of posts these days. Japan's lovely and I'm out every day. Additionally my family back home can't into rebooting a downed server so I've been struggling to get Schwartzwelt back up. Though, Mitra's been advancing steadily so if I can have foreign images baked in instead of mediaproxying I'm moving everything here.
Connecting to my servers now is slow as heck. Anyone got a decent guide on CDNmaxxing for Pleromer/Mitra without relying on Cuckflare? Main options I'm aware of are Bunny CDN, S3 buckets and having even moar VPS'.
@silverpill I've noticed relays ending in actor (e.g. https://relay.poster.place/actor and https://relay.fedi.agency/actor) don't show up whatsoever. Most Pleroma instances I notice have a relay actor I can find but it's inconsistent and my @relaystalker seems to be stuck at follow requesting for the ones I tried to follow. If I missed something obvious please let me know!