In the early days of the web, monitoring availability was simple: a server either responded to a ping, or it didn’t. HTTP checks tightened that up a little — a 200 OK meant the dashboard turned green, and everyone assumed things were fine. That assumption doesn’t really hold anymore, though.