We've analyzed the apps run on the affected machine and none of them appear to have had any IP-whitelisting-related errors. We have restored the cluster to its normal state.
Posted Aug 03, 2017 - 19:39 EDT
A small number of US apps with IP whitelisting enabled were run on a machine without IP whitelisting configured for the 20 minute period between 16:07 and 16:27 Pacific time. We are restarting all affected containers on the correct machine now.