This incident has been resolved.

There was a 16 minute period where one of the web servers in our cluster was raising errors if you happened to hit that server. We did a reset and the errors have cleared. We will continue to monitor.