We are currently investigating this issue (status.claude.com)

I am genuinely currious. With such terrible uptime... is it really just a series of "oops, we need to fix that" or is it that they just keep hitting a capacity wall and have to drop service until they add capacity or reduce usage?