When the dust settles, I'd be curious to know whether some unexpected and accidental synchronization occurred that put EC2 into it's mangled state.