I've had an AWS instance running a web & mail server for 6 months. It crashes often, about once per week. It usually takes 15 mins - 2 hours before the server is up again. One time I contacted Amazon about this and they gave me a stock response which essentially said it was a hardware failure. I don't believe my instance runs into hardware failures once a week.
What steps should I take to find out why my instance has so much downtime?