At the KubeCon 2017 event in Austin, Texas, AutoScalr announced integrated support for Kubernetes. Kubernetes is one of the most popular Docker container orchestration solutions in use today, and with the announcement last week from AWS at re:Invent for a new managed Kubernetes service (EKS), the Kubernetes adoption momentum seems to only be accelerating.
The AutoScalr integration allows leveraging the AutoScalr machine learning autoscaling technology to run a Kubernetes cluster on AWS at a fraction of the normal cost, between 50-75% less, by automatically selecting the most cost-effective set of instances types and blending in the use of spot instances diversified over many spot markets that also meet the availability requirements of the cluster.
The integration is similar in concept to the recently announced AutoScalr integration with ECS, but leverages the native autoscaler component that is already a part of the Kubernetes project to make it quicker and easier for existing Kubernetes clusters to take advantage of the AutoScalr technology for cost savings. The integration can be enabled in just a few minutes by making three simple changes to your existing autoscaler yaml deployment file. The integration is open source, and can be found on GitHub along with further documentation at: https://github.com/autoscalr/k8s_autoscalr