We have resolved an issue with our core message processing functionality in the US cluster. We are monitoring the service at this moment and will conduct a postmortem shortly.