This was a result of us throttling traffic to the site - not the original outage.