Hard Problems We Handle in Incidents but Aren't Recognized. John Allspaw @ SRECon21. Experts, like the audience of engineers at SRECon, are so good at responding to incidents, it makes it really hard to see how hard the work actually is. Allspaw outlines four categories of cognitive activity and two or three specific challenges. program
youtube ![]()
Four groups of cognitive activity during incidents: diagnostic 3:29
, therapeutic 4:34
, recruiting 5:59
, reporting 7:29
.
Couple costs of coordination 10:06
. See Dr. Laura Maguire's dissertation or SRECon talk from last year: Incident Cognitive Costs
Sacrifice Decisions 18:52
See Stella Report: Human Performance in Systems
Parallel Incidents 22:43 ![]()
Optimistic Applause 27:51 ![]()
We are way better at this stuff than we think we are.
It's difficult for us to see what makes us good at it.
Right here's the thing. The most beautiful, amazing, incredible thing. You are why your services stay up. Why things actually work and stay working.
You all are the sources of resilience.
It's not the load balancers. It's not the designs. It's not the distributed blah blah blah.
Absolutely make no mistake it is you. It's all of you.
The hardest challenge is understanding how you do what you do. Your expertise. 29:14 ![]()
"We know more than we can tell." — Michael Baliani, 1966, talking about tacit knowledge.
.
Although the whole talk avoids the jargon, this is entirely a story pulling the curtain back on the Law of Fluency.