This also happens on Red Hat. It will recover in like 30 minutes even if you do nothing. Guess this is why aws sucks.