The root cause of this incident is an emergency code release intended to gather new metrics to help monitor and further mitigate the recent incidents we have been experiencing in our European data center. The monitoring code performed well in testing, however, it could not keep up with production throughput. Our engineers rolled back the change and are reworking the metrics capture update for the next release.