On Being Wrong (Again) – Now With Metrics
Recently, I wrote about being wrong. Specifically germane to this post, I wrote: Being wrong means I probably had some gross misunderstanding of a system’s architecture or the operation of a program, and that means I have an opportunity to learn more about it, and hopefully to be able to guide my decisions. I was involved in an incident where I thought that some EC2 instances had reached disk saturation. This was primarily driven by three factors: referencing an outdated man page, tunnel vision, and missing the units on a graph....