Getting Creative with HPA Scaling in Kubernetes
summary
Demo compares VM ASG that takes ~6 minutes to bring capacity vs Kubernetes pods that scale in seconds and shows HPA replaced by external/custom metrics via Prometheus and KEDA. Examples use one pod per 50 requests for an Ingress controller, a viral-prep rule that ups 10 pods per 50 Twitter mentions, and a Twitch flow that keeps replica 0 until stream starts then adds one pod per 100 active users; all examples run live and scale as expected.
AI-written summary · about
related talks
30mHorizontal Pod Autoscaler Reloaded - Scale on Custom Metrics
51mHorizontal Autoscaling with Kubernetes
29mKubernetes Autoscaling: Beyond One Size Fits All
21mMusic Matched to Different Ways to Scale in K8s, What Could Go Wrong?
47mKEDA, HPA and VPA? An Introduction to Scaling
13mUsing Virtual Nodes to Auto-Scale Your Application on Kubernetes
34mAutoscale your Kubernetes Workload with Prometheus - Frederic Branczyk, CoreOS
32mScaling Kubernetes Application to 1M Users