During a routine network upgrade, there was a brief moment between 11:00-12:00 UTC in which traffic was erroneously routed, causing some unintentional downtime for users using custom domains and subdomains. The issue has been identified and remediated.