Kevin Sookocheff

I was recently introduced to a simple article on system design: Users are almost always right. In it, a simple rule is stated: This rule reminded me of the concept of desire paths: unplanned trails created by human or animal traffic that usually represent the shortest or easiest route between two places. Desire paths in an urban setting often run counter to what a planner might suggest, leading to a natural conflict between the pedestrian and the planner.

Application Portfolio Management (APM) draws inspiration from financial portfolio management, which has been around since at least the 1970s. By looking at all applications and services in the organization and analyzing their costs and benefits, you can determine the most effective way to manage them as part of a larger overall strategy. This allows the architect or engineering leader to take a more strategic approach to managing their application portfolio backed by data.

Shortly after deploying a new Kubernetes cluster, one of the first things you will likely want to do is collect some metrics and data about how it operates. There are two projects that are typically used for this, and since they are named similarly it can be confusing to know which one you should use and why. This post hopes to clear up any confusion between the Kubernetes Metrics Server and kube-state-metrics.