Some remarks on the size of tubular neighborhoods in contact topology and fillability