Awesome article with great detail and explanation. I've always been curious about how load averages are calculated and what exactly they mean.
The average number of threads/processes in the run queue and/or blocked on disk i/o sampled at 1, 5, and 15 minutes.