Kubernetes: Autoscaling plus Karpenter does not equal optimization
Kubernetes optimization
Did you know autoscaling can increase costs if workloads haven’t been optimized?
There is a common thought that when you implement technologies like HPA and Karpenter, you are also taking care of optimization. The reality is that autoscaling is not optimizing. It is important to allow for scaling up and down but in the absence of resource optimization, you will never be in the most cost effective state. In fact, autoscaling can actually make you less efficient than with a more static configuration.
In this 30 min session you will learn about strategies that are the best of both worlds; we cover 3 reasons autoscaling does not equal optimization:
- In most cases autoscalers don’t optimize pod requests and limits
- Autoscaling unoptimized pods or nodes simply amplifies the issue
- Even autoscaling solutions that optimize container sizes have opaque algorithms and fail to account for historical trends
Who would this benefit?
- FinOps Practitioners with responsibility for K8s
- GKE, AKS, EKS Platform Owners
- OpenShift Platform Owners
- Cloud Platform Owners
- SRE’s concerned about cost
Interested in seeing Densify in action? Request a demo »
About the Presenters
David Chase Technical Specialist, Densify
David Chase is a Technical Specialist at Densify focused on Cloud and Container Resource Control & Optimization. He has over two decades of experience helping enterprise customers solve business problems with software and technology. He attended Sheridan College where he studied Computer Science.
Andy Walton Americas Sales, Densify
Andy Walton has over 25 years of experience in IT. At Densify, he works with new prospects and business partners to demonstrate the unique value and functionality of the Densify service. Andy has a bachelor degree with honours in mathematics and computer science from the University of Waterloo.
AVAILABLE ON
PROUD MEMBER OF
