It'll be interesting to see how this works out. But I think the cause was DNS load balancing and a race condition that caused the issue. Amazing how DNS can cause such a huge outage.