Conversation

Notices

1.  (mint@ryona.agency)'s status on Monday, 05-Aug-2024 04:39:43 JST 

   in reply to

   • Hoss Delgado
   • :blank:
   • Angry Sun
   @i @meowski @Hoss @sun Mediaproxy could be decoupled from the instance by setting up a secondary pleroma installation with the same secret_key_base in config. That said, it never touches postgres so it shouldn't impact DB performance to begin with.
   • † top dog :pedomustdie: likes this.
   • :blank: (i@declin.eu)'s status on Monday, 05-Aug-2024 04:39:44 JST :blank:

     in reply to

     • Hoss Delgado
     • Angry Sun
     @sun @meowski @Hoss with logging it should be simple enough to see, could just be a thunderherd of cdn pulls stemming from the mediaproxy
     
     i know for sure pinning three movies worth wasted a fair bit of network traffic
   • Hoss Delgado (hoss@shitpost.cloud)'s status on Monday, 05-Aug-2024 04:39:45 JST Hoss Delgado

     in reply to

     • Angry Sun
     So DB rot isn't an unavoidable inevitability of long-lived Pleroma instances?
   • Angry Sun (sun@shitposter.world)'s status on Monday, 05-Aug-2024 04:39:45 JST Angry Sun

     in reply to

     • Hoss Delgado
     @Hoss @meowski it definitely got slower and slower over time but it wasn't cripplingly slow all the time, only when the network traffic spiked
   • Angry Sun (sun@shitposter.world)'s status on Monday, 05-Aug-2024 04:39:46 JST Angry Sun
     @meowski the two problems were excessive DB access and excessive bandwidth usage. They were probably connected and if I figured out what the root cause was it would fix everything