Between approximately 8:37 AM and 12:01 PM US Pacific time on 8 March, named user changes and named user tag changes on our US cloud site may have been delayed in excess of 1 hour, up to a maximum of 2.5 hours. We also experienced some short periods of elevated 5xx errors from named user API endpoints as the system recovered towards the end of this period. While this should not have impacted message delivery directly, it's possible that audience lookups for sends may have utilized data that were not yet up to date. This issue was due to elevated levels of a particular kind of traffic, and was resolved by blocking that traffic and reaching out to the customers involved to help them use our platform more efficiently. Our Engineering team is investigating changes we can make to the system to handle this sort of traffic in large volumes better in the future, preventing future recurrences. We apologize for any inconvenience.