This incident has been resolved.

A fix has been implementing, things have appeared stable since 12:30pm pacific time, we'll continue to monitor the results.

Seems we haven’t solved it fully. We’re investigating intermittent response time and dropped requests again.