@john_weigand, we actually want to alert well before autogrow runs out so we can resize a volume before it reaches the 'end of the line'. Our philosophy is - we rely on autogrow to help fix the problem for a few minutes if we can't get to an alert immediately, but we want to resize the volume well before it reaches the autogrow cap.
With that in mind, for volumes under 500 GB we have alerts set to 68% and 76%, which equals out to 82% and 92%. For volumes over 500 GB they are set to 76% and 80%, which equals out to 92% and 96%. The fact that we have to do it this way is ridiculous to me, and confusing to those who manage alerts. They say, "why am I getting an alert when it's only 76% full?", when it's actually 92% full.