Date: 08.29.2024<!--date as MM.DD.YYYY-->
Location:Ga Tech ATDC
Speakers:
John Platt, CTO @ StormForge <!--presenter name @ company-->John has a PHD in Applied Maths and has spent most of his career building Machine Learning powered products. Throughout his career he constantly found himself using machine learning to solve infrastructure related problems that he didn’t want to toil with. The desire to automate away mundane tasks combined with a desire to solve the hard problems with maths is what led him to his current role as CTO of StormForge, which focuses on Kubernetes resource optimization.Kubernetes Autoscaling: Beyond One-Size-Fits-All<!--presentation title-->There’s no one-size-fits-all approach to Kubernetes autoscaling. Fitting the right pieces together requires a deep understanding of the various autoscaling dimensions and the open-source projects that can plug in to meet specific use cases.This session will explore the three key dimensions: cluster, horizontal, and vertical. We’ll start with the tools built into Kubernetes before moving onto the more advanced open-source projects—Karpenter and KEDA, which enhance flexibility and efficiency for each.Attendees will learn:
Strategies for cluster, horizontal and vertical autoscaling, their challenges, and solutions that improve on or replace them
How to assess workload requirements to choose the right mix of autoscaling tailored to specific app needs
Best practices to optimize deployments for peak performance and cost-efficiency for dynamic workloads