We can define volatility (per white's position or black's) roughly by some function of the number of moves which result in a similar (or better) positional evaluation.
For example, in the fried liver attack, white's volatility is low since white can make a lot of fairly natural looking moves (whether they add to the attack it doesn't matter) which result in no significant loss in the evaluation of the position. With black in the same attack, the volatility is high since black has to make very accurate moves in order to maintain equality or a better position.
I haven't seen any engine measuring this volatility, but I think it is important because you can mislead your opponent into thinking he has a better position but if his position is highly volatile, you have many chances to take advantage.
For example, in the fried liver attack, white's volatility is low since white can make a lot of fairly natural looking moves (whether they add to the attack it doesn't matter) which result in no significant loss in the evaluation of the position. With black in the same attack, the volatility is high since black has to make very accurate moves in order to maintain equality or a better position.
I haven't seen any engine measuring this volatility, but I think it is important because you can mislead your opponent into thinking he has a better position but if his position is highly volatile, you have many chances to take advantage.