I've seen literally dozens of unresponsive and defective EC2 instances over the last year. This was spawning 100s of medium and large instances per day, with an average instance lifetime of around 3 hours.
From what I've read during my usage of AWS over the last couple of years, this is more the norm than the exception.
I think we agree then. Dozens out of 36,000+ instances is a very low failure rate. I haven't had nearly that volume so I've only experienced 2 failures on EC2.
From what I've read during my usage of AWS over the last couple of years, this is more the norm than the exception.