Recently, Amazon had a major outage with its Amazon Web Services (AWS) system. Thousands of companies were affected for hours. The problem showcased which companies had their own infrastructure and which were simply cloud-based.
What surprised me was the lack of a plan when things go wrong. Certainly, that starts with Amazon, but it also cascades down through the entire chain. In my case, that meant how the approval process was accomplished.
I’ve used third-party companies to host approval clips for years. It’s convenient and includes features for adding comments and posting versions. It’s also cost-effective and eliminates the need to maintain an infrastructure locally.
Over the years I’ve used various services for this. And I’ve learned that while it’s convenient, it’s not always available. To say something is up 99 percent of the time means that it could be down for more than three days. Even 99.9 percent could still be out for more than 8 hours.
That doesn’t sound like much, but if those 8 hours are when you’re trying to make a deadline or merely want to get something posted so you can go home, it can be frustrating. So that’s why I always have a plan B.
I’ll admit that I expect third-party companies to have a plan B. I’ll also admit that I was surprised how many companies relied entirely on AWS for their infrastructure—or so it seemed. And while the AWS outage was finally fixed in about 17 hours, just think about that: 17 hours. It took even longer for the backlog of work to process at third-party companies. In many cases, work had to be re-uploaded/restarted.
Of course, this was right in the middle of a project I was working on that had a tight deadline. (Outages never happen when you aren’t pressed for time.) But I had my plan B, a local server I had access to for posting and linking to files. It wasn’t as pretty and convenient, but the work got in front of the eyes that needed to see it.
If that didn’t work, I had a plan C, too.