I've been studying cognitive work in software environments for some years now. I want to bring attention to a phenomenon that is both familiar (in an "uh, of course!") and AFAICT does not receive much attention. 1/n
Before I write about that, I'll mention that a primary way this research is done is by using a family of methods called "process tracing" which involves the triangulation of multiple data sources to make analytically valid inferences about cognition "in-the-wild." 2/n
In a nutshell, we take behavioral data (what people actually *do*, even if they don't say they do it) and what they externalize (verbal, textual, etc.) to inform further exploration of the cognitive processes. It's obviously much more than this, but that's the gist. 3/n
Point is: cognitive processing requires *deep* investigation and exploration. Anyway, a phenomenon is that when faced with uncertainty or ambiguity in responding to anomalies with time pressure (such as outages, etc.) people are agnostic about what tools they use for help. 4/n
If command-line is what they think will help, they'll use it. Same with logs, dashboards, alerts, conversation with others, wiki pages, blog posts...anything. In those circumstances: what they believe helps has value - has value, full stop. 5/n
People in these scenarios do not make distinctions between observability/monitoring/time-series/telemetry/oldschool/newschool/elegant/hacky tooling; the *only* criteria is if they believe it will help them. Where those beliefs come from is complicated and a different story. 6/n
A secondary but just as important finding is in these time-pressured scenarios, people will not hesitate to "drop-down" layers of abstraction wrt to tools if they think that getting "closer" to the "metal" will help them or they have any distrust in the higher layer ones. 7/n
In other words, if people don't entirely understand what they're seeing in GUI tools, they'll reach for tcpdump/gdb/dtrace/etc. in a fluid, almost unconscious fashion. This is referred to in the Stella Report (snafucatchers.github.io/#3_4_2_Basic_t…), but needs greater attention. 8/n
If I were building GUI/higher-level tooling to help people in these situations, I'd devote a great deal of time and attention to this. Fin.
