Session Name: Karpenter: The New Kubernetes Autoscaler Focused on Node Optimization
In this session, you will learn how to provision, manage, and operate Kubernetes clusters at any scale with Karpenter. Karpenter is a Kubernetes cluster node lifecycle management solution. Karpenter monitors the pending pods on schedule and launches the necessary instances efficiently. Instance selection decisions are driven by the pods specification, including the resources it requests and scheduling constraints. I'll use EKS to run a small pool of nodes and deploy a minimal set of On-Demand instances. Next, I'll use Karpenter to deploy a mix of On-Demand and Spot instances to show some of the benefits of running an autoscaler without autoscaling groups (ASGs). Karpenter + EC2 Spot Instances allow you to run large-scale and low-cost optimizations workloads on Kubernetes.
Speaker Bio:
Christian helps customers to build fault-tolerant, elastic, reliable, and cost optimized workloads in AWS. He’s passionate about Kubernetes, programming, and building tech communities. Christian has spent 15 years working in different companies to build modern solutions using the cloud.