A few days ago, I talked about the value of knowing what your APIs are doing. Today, I wanted to use a TOTALLY hypothetical example (that happened to me) on an API fail.
I use a video streaming company. Pretty cool company, got their start in DVD rentals. They switched to a microservice architecture a few years back, meaning the service essentially never goes down. Essentially. But let's walk back a couple of years ago. My wife is sitting next to me, and decides to watch an episode of one of the series she follows. She logs into said video streaming company on her iPad, pulls up the episode, and proceeds to tear up over what's happening in this particular series. Being a caveman, I decided to watch something else. I pulled out my trusty Nexus 10, and pulled up the app to the same video streaming company, and pulled up a very very different episode of a very different series. But....when I selected the episode, nothing happened. While the site was up and running, and the microservices were doing their job...essentially....there was an API failing somewhere in the chain, but only for Nexus 10 users. Meaning, my wife could tear up, and I could...well, read a book.
With API monitoring, this fail may still have happened, but an administrator would have been notified immediately, and the outage would have lasted seconds instead of hours. And while hypothetical (cough), one could imagine if this kind of fail had happened in a banking transaction....a retail transaction....or even an insurance transaction. API monitoring should be considered mandatory for mission critical apps, no question.
I'd like to invite you to an upcoming webcast, One bad API(L) Can Ruin the Entire User Experience, taking place tomorrow (September 26th at 9am (PST)). Brian Whitmarsh, Sr. Director, Product Management for CA Technologies Agile Operations Business Unit, will be discussing API monitoring in depth, and showcase Runscope, an API monitoring solution that's tightly integrated with CA API Management.
I look forward to seeing you there!