There is a tendency to imagine (or remember!) incidents as unfolding much neater and orderly than they actually are. Events can lead some engineers scratching their heads about what is happening, while their teammates can instead be confused about how it’s happening. Sometimes diagnostic paths can lead down rabbit holes. Sometimes issues can be detected after they’ve been resolved. This talk is about recent research on these “messy details” of real incidents and it’s implications on how we understand them beyond the traditional “metricification” approaches.
John Allspaw, Founder, Adaptive Capacity Labs LLC