API downtime: Who’s asleep at the wheel?

Many popular applications use the application programming interfaces (APIs) of other services, from your favorite Twitter client to Facebook apps. So downtime on the most popular APIs has a serious knock-on effect on other services and (up)time is money.

Performance monitoring company WatchMouse just published a study on the downtimes of the 50 most popular APIs including Google Maps, Flickr, Twitter, YouTube, Amazon, eBay, Facebook and Wikipedia.

[aditude-amp id="flyingcarpet" targeting='{"env":"staging","page_type":"article","post_id":251014,"post_type":"story","post_chan":"none","tags":null,"ai":false,"category":"none","all_categories":"business,","session":"B"}']

WatchMouse tracked these APIs between February 16 and March 17. The APIs were tested every five minutes by authenticating and calling a single operation on the API (rather than all of them). In general, if one operation on an API is available, they all will be.

Ten APIs, many of them from Google, had 100 percent uptime including Basecamp, eBay, Google Charts, Google Maps, Google Search and Quora. Availability of less than 99 percent is regarded as “poor” since this translates into around 1 business day of downtime per month. Imagine if Internet access was unavailable in your office one working day a month.

The worst performers, managing less than 99 percent availability, were Digg, Gowalla, Geonames, Eventful, and Posterous, with MySpace coming in last with an uptime of 94.32 percent. This means two days of downtime per month. Paypal and Salesforce also didn’t do well, struggling to keep above 99 percent availability.