Network issues with www and api

Rackspace/Slicehost appears to have a network routing issue, and our machines in their network are not routed correctly in some parts of the world. We are investigating the issue and will update this with more information.

Update: The problem was external to Rackspace/Slicehost, however it caused our main validation server to lose connectivity and queue up OTP sync requests to our other servers. The system recovered when connectivity was restored (as it is designed to), however the database insert/delete performance remained poor even though the queue table was near empty. We are investigating how this can be avoided in the future, and will likely require some software tuning.