If you are going to build an application on Facebook, you’d expect that users will be able to access your application as long as it’s developed properly, but unfortunately Facebook’s reliability has not been at its peak recently. Yesterday the company posted that they were “investigating issues that may be causing some Platform apps to fail to load”. While the volume of errors appeared to hit a new high yesterday, the Facebook Platform’s reliability had been at an all-time low for over a month now.
While speaking at the Facebook Developer Garage at SXSW last week, Gareth Davis, Facebook’s program manager for games, fielded a question from a developer who asked about the Platform’s reliability. At the time he admitted that “February was a difficult month”, according to a number of attendees. It appears that the company is continuing to experience some challenges in maintaining Platform reliability this month as well.
Today the company was dealing with API latency issues which was a continuation of some of the problems from yesterday. One developer told me off the record that he believed Facebook was cutting down on app virality for two reasons: increase ad spend by developers (as advertising is an integral component of successful applications) and to increase the platform reliability.
So far it appears as though the initial drop in traffic has only affected specific categories of applications (dating and quizzes), but the larger issue for most developers has been uptime. Facebook is working hard to improve the reliability of the platform, but scaling a Platform which serves over 400 million users while opening up the information available to developers has been difficult.
While Facebook has not specifically articulated the causes of the issues the company continues to work on improving API reliability as it’s not only critical to the success of the Platform but Facebook Connect as well. With Facebook preparing to open up even more data at the upcoming f8 event, it’s important that the company resolve the Platform uptime issues sooner rather than later.