Cloud Native Prague #7

Scaling down your unused services during night or weekends is a common practice nowadays. But what happens when someone actually needs them? How can they turn them on with the least amount of hassle? Let's find out.

One of key characteristics of modern platforms is their ability to dynamically scale applications, including a scale down to zero when the application is not used so that they don't cost anything. This session gives an overview of how Kubernetes, the leading container orchestration platform, allows elastic scaling of your application. You will learn about Kubernetes intrinsic auto-scaling capabilities like the Horizontal Pod Autoscaler (HPA) or the Vertical Pod Autoscaler (VPA) and how to use it with your application. Kubernetes does not support autoscale-to-zero or custom metrics out of the box. But luckily, there are projects like KEDA that brings those missing capabilities and more to your cluster. We will learn about the right approaches to autoscale apps and see live demos.