Kubernetes Cost Optimization with KEDA

Transcript

1. SPONSORS Speaker: Arek Borucki Company: Beamery Mastering Kubernetes Cost Optimization with KEDA and MongoDB
4. • Autoscaling • KEDA architecture • Scalers • RabbitMQ, MongoDB • Demo Agenda
5. LET’S GET STARTED
6. Autoscaling • Dynamically adjust resources • Reduce Kubernetes Costs • K8s built-in autoscaling mechanisms 1. HPA - increase or decrease replicas 2. Cluster autoscaling 3. VPA - Recommender
7. KEDA • Kubernetes Event Driven Autoscaling • CNCF incubation project • scaling containers based on the trigger • allows scaling to/from zero • K8s operator - add into any K8s cluster • KEDA works alongside with HPA
9. CRD
11. if deployment should be activated or deactivated
16. Job Application
20. SUMMARY • Autoscaling significant reduce cost! • HPA, VPA, Cluster Autoscaler • KEDA can use custom metrics for a specific event source. • Create an index on query pattern SUMMARY • Autoscaling reduces cost • HPA, VPA, Cluster Autoscaler • KEDA uses custom metrics for a specific event source • KEDA is pro-active and dynamic • Create an index on query pattern
21. SUMMARY • Autoscaling significant reduce cost! • HPA, VPA, Cluster Autoscaler • KEDA can use custom metrics for a specific event source. • Create an index on query pattern SUMMARY • Autoscaling significant reduce cost! • HPA, VPA, Cluster Autoscaler • KEDA use custom metrics for a specific event source. • KEDA is pro-active and dynamic • Create an index on query pattern THANK YOU