>
We had all kinds of downtimes due to AWS over the years.No, you just had downtime, full stop. Failures are inevitable, regardless of whose datacenter you're using. Failure-tolerant architecture design minimizes (or eliminates, if you're lucky) the downtime caused by failures. I'm not saying that's easy or necessarily appropriate depending on the stage of your company, but foisting responsibility onto AWS is missing the point.