Twitter had issues early on, but before Musk bought them they definitely did not have outages half the time.
I've become accustomed to the fact that it's beyond clear that pre-Musk engineers pushed changes to the servers live without testing them first, and didn't even bother doing an incremental roll-out either. Either they had no engineering culture, or that culture was gutted long before he arrived, because Twitter had problems with frightening regularity.