A “race condition” after the system was reset prevented the sequencers from catching up, causing the second outage.