> Expect failures; what's more, embrace them I appreciate the point he's driving home, but his rate of failure is meaningless without knowing the size and upti

> Expect failures; what's more, embrace them I appreciate the point he's driving home, but his rate of failure is meaningless without knowing the size and uptime of his "server farm". Is it 10 servers or 100 servers? Have they been up for a month or for a year? I pull this point out specifically because I personally had 4 EC2 instances epically fail on me over the course of 18 months and 40 instances. Of course, this was in year 2 of EC2, so I'm sure things are better now, but I believe it's an important data point.