Dynamic Scaling in Kubernetes: Best Practices and Tools for Success

Kubernetes has become the de facto standard for building and deploying modern applications. However, as teams migrate their projects to Kubernetes, they often overlook critical factors such as deployment environments and technical constraints. Most of our community operates managed Kubernetes clusters, with AWS EKS providing comprehensive tooling to smoothly provision and manage these clusters. Have you encountered limited access to observability data for the Horizontal Pod Autoscaler (HPA) or faced unnecessary costs due to resource provisioning? Organizations typically create autoscaling rules based on CPU and memory usage, but ensuring reliability and cost efficiency solely with CPU metrics can be challenging in a Kubernetes environment. How can we effectively leverage observability data to improve our autoscaling policies? Discover: - Karpenter: An open-source solution that helps you efficiently provision your Kubernetes clusters. - Dynamic Scaling of Observability Data: Learn how to scale your observability data dynamically. - Automated Decision-Making with Dynatrace Predictive Algorithms: See how you can automate your decisions using advanced algorithms. Join our presentation to see how the AWS EKS and Dynatrace can simplify autoscaling with the Horizontal Pod Autoscaler. Through a real-world example showcasing inefficient HPA rules, we will demonstrate how to enhance and optimize the autoscaling process using metrics from Istio/Envoy, while also considering the cost of your workload.