There is a vast amount of literature about running Kubernetes in production but… when there are problems? how can we quickly diagnose what is going on? Kubernetes is like a black box which usually takes few disasters to get the grasp on.
In this talk we are going to learn how to use monitoring in our favour not only for detecting issues but also to predict them and save time and money: When is our hard drive is going to be full? How long will take with the current growth to need more nodes on the cluster?
These and more questions will all be solved in the talk with examples and live demos showing to the audience how a strong monitoring culture will help your company to save a lot of time and stress.
We will deploy a new version of a running app and explain what happens at the monitoring level and how we can detect a botched release even before than our full customer base.