For quite a while now I've been trying to spread the idea that monitoring should be done top-down and not bottom-up. The idea in a nutshell is this: Stop measuring everything and trying to deduce overall system health from low-level metrics; instead start by defining your business KPIs and only then decide what you need to measure and what's the best way to measure that thing. This idea corresponds well with DevOps culture, since in most cases it would be up to your developers to instrument their applications to send the relevant metrics, thus breaking the dichotomy between dev and ops.
But recently I've realized that it's not enough to just involve the dev team. After all in most organizations it's not the dev team's job to define which KPIs are critical to the company's business, right? At the least - product managers should be involved, but at best it's a cross company effort involving sales, marketing, and even the CEO! So I concluded that "DevOps" is not enough. It doesn't capture the size of the effort that companies need to invest in order to do operations the right way. So - with some help from Miley Cyrus (yes, the pop star) - I decided to coin a new term: PanOps. In my lecture I intend to talk about what PanOps is, why I think it's important, and how it manifests itself in areas such as monitoring and automation.
Father, husband, thinker. Passionate about cinema, philosophy and building distributed systems.