Yesterday I highlighted a platform “bug” that was the result of a poorly designed script. Last night, Facebook posted an update to the platform status feed:

The current update on the requests/invites is that it looks like the requests that were sent out over the course of the day on Monday are lost. We are working on recovering requests that were sent out up until Sunday, those should hopefully be recovered after tonight’s push.

Well guys, it looks like we’ve slipped up a little and have lost millions of application requests. Sorry! As if the errors weren’t enough, the platform went down last night from about 12 AM until about 2 AM Pacific Standard Time causing trouble for thousands of Facebook developers and their applications. While it’s understandable that most changes are pushed at night during the lowest traffic time, Facebook should have a better testing environment to have developers that can test changes during peak hours.  If Facebook already has a good testing environment, they should use it!

Last night I had a number of developers contact me to get across how significant a snafoo this was. Most of these people have their entire businesses built on the Facebook platform and generate their revenue from traffic coming to their applications. For applications that have millions of users, tens of thousands if not hundreds of thousands of requests could have been deleted. That has a significant impact on those applications’ traffic. Was you application affected by this mishap?