On May 15 at 1:48 PM ET we experienced a brief database failure due to a cleanup process that didn't complete successfully. This, in turn, caused all outbound and inbound messages to be delayed for a little over an hour (a small number of messages were delayed for a longer period of time). We were still accepting and processing messages throughout this incident, and all messages were eventually delivered once the database issue was resolved.
We pride ourselves on performance and speed to the inbox, and this is a clear failure on that commitment to our customers. We are doing an internal audit of our alerting and monitoring as well as a review of the infrastructure that queues messages. Our goal is to make sure you can consistently rely on us for the best service possible and we will work hard to make sure this does not happen again.