man people acting like you gotta use elasticsearch to search toots
okay, using a LIKE query on the amount of toots an average server has is fucking insanity and it's /slow/
but ES is like using a piledriver to pound in a nail
lucene would have been sufficient, and lucene isn't all that bad in my experience to run
Wikipedia's been using something like this for eternity and it seems to work just fine for them
(yeah, Wikipedia scale, Masto is nowhere near that)