> I'm also intrigued by the constant cloud refrain of "stuff crashes all the time so just expect it to" coming from a background where I have apps that run with

> I'm also intrigued by the constant cloud refrain of "stuff crashes all the time so just expect it to" coming from a background where I have apps that run without crash for 6 months at a time, or essentially until the next release. IMHO, those aren't mutually exclusive. Your app code should be robust enough to run 6+ months at at time, and the "stuff crashes all the time so just expect it to" attitude should be reserved for stuff outside your control, like hardware failures.