This was a one-off issue caused by a data inconsistency introduced during the development of a recent feature. We’ve fixed the inconsistency and added automated checks on the services that depend on that data so that this issue doesn’t occur again.