When you're running a world wide service with lots of moving and connected parts this dramatically increases the danger you'll have catastrophes that will cost much or most of that market share.
As a thought experiment, suppose all of AWS was down for a full week? What would happen after that?
More realistically, how many times can us-east-chaos-monkey bring down various global services before a lot of its users start making moves to mitigate or eliminate their exposure to AWS?
For a more concrete example, I'm working on a project that has an easy but critical use of cloud services. The more I read about Amazon following Microsoft into Ballmer era stack ranking madness, and the most AWS fails objectively fails, the lower the priority it gets for which cloud platform to support first and best.
As a thought experiment, suppose all of AWS was down for a full week? What would happen after that?
More realistically, how many times can us-east-chaos-monkey bring down various global services before a lot of its users start making moves to mitigate or eliminate their exposure to AWS?
For a more concrete example, I'm working on a project that has an easy but critical use of cloud services. The more I read about Amazon following Microsoft into Ballmer era stack ranking madness, and the most AWS fails objectively fails, the lower the priority it gets for which cloud platform to support first and best.