Application Autoscaling Made Easy With Kubernetes Event-Driven Autoscaling
summary
Kubernetes Event-Driven Autoscaling (KEDA) is a controller and set of CRDs that lets existing Deployments and Jobs autoscale on external systems (Kafka, Redis, Prometheus, Azure/AWS/GCP) without wiring metric adapters. Install via Helm or operator, create a ScaledObject or ScaledJob plus trigger authentication, and KEDA uses an HPA under the hood to scale to zero and back. Demo: an order processor handled 250 queued messages, scaled from 1 toward a configured max of 25, then dropped to zero when the queue emptied. HTTP support is alpha via an interceptor.
AI-written summary · about
chapters (6)
- 0:00Start
- 2:25Application autoscaling made simple with Kubernetes Event-driven Autoscaling (KEDA)
- 5:35How does it work?
- 7:05Production-grade authentication
- 21:47First-class support for HTTP-based
- 24:55Roadmap
related talks
42mSmart scaling with Keda
29mKubernetes Autoscaling: Beyond One Size Fits All
36mAutoscaling Isn't Set-and-Forget: Tuning KEDA & Karpenter for Real Workloads.
46mKarpenter * Keda: The Winning FinOps Duo
42mMaking Resilient, Self-Adaptive, And Autonomous Kubernetes Clusters
50mAutomating Kubernetes with Event Driven Ansible
28mScaling in Kubernetes Safely on On-Prem KaaS Across 1,300+ Clusters and 40,000+ Nodes
57mScaling Azure Container Apps: Maximize Performance and Minimize Costs