Fixed.
Damn failover setup is quite sub-par and doesn't automatically fail-back to the main server. As such, I have to manually revert back to this one... it appears I did not switch WWW back to this server... If you ping it, you'll see that went to a Ramnode box in NYC. vpsBoard is hosted on an BuyVM box in NJ.
StatusCake and Rage4, source of much frustration right now. It will failover automatically to the failover (RamNode) server if I shut down nginx, turn off the server, if vpsBoard is getting hit with a large flood that isn't filtered instantly, or if there are upstream issues. That's great, and it works as intended (though slow, since StatusCake only checks the availability once every 5 minutes.)
The issue is, it doesn't go back to the BuyVM server once it is reachable again. I tried several different configurations and it's something I'll figure out this weekend.