Amazon EC2 suffers major downtime
A large portion of Amazon’s EC2 cloud-computing service is having big problems today, causing downtime for a lot of popular websites that rely on it. As of now, it’s been intermittent or down for almost 7 hours. (Status.)
It might be easy to say, “Don’t host on EC2 because things like this happen.” But there’s no such thing as perfect uptime.
Rather than dwelling on specific failures, which all hosts and platforms will have, you should ensure that you have a viable contingency plan should you need to change hosts.
If you’re hosting your service (or a nontrivial portion of it) on EC2 or other managed cloud services, how easy is it to move elsewhere? How orthogonal is your host to your service? If your cloud service suddenly shut down and you could no longer use it, how much downtime or data loss would you suffer before restoring full service?