Sweet post. I think despite the massive AWS outage last year, a lot of setups still aren't as fault resistant as you'd want. Last year's AWS outage died cuz of a cascade failure on EBS. Have they resolved that issue or should your EC2 volumes not be on EBS on this setup?