Overview
Dive deep into the world of autoscaling in Kubernetes with this comprehensive conference talk by Maciek Pytel and Joe Burnett from Google. Explore the intricacies of Horizontal Pod Autoscaler (HPA), Vertical Pod Autoscaler (VPA), and Cluster Autoscaler as members of SIG-Autoscaling reveal their inner workings. Learn how to debug and optimize autoscalers to reduce infrastructure costs while balancing availability. Discover where to find crucial information about autoscaler activity, which settings can be adjusted, and which flags are best left untouched. Gain insights into the recommendation lifecycle, custom metrics implementation, stabilization techniques, and backwards compatibility considerations. Examine the Cluster Autoscaler architecture, scale-up processes, and expander functionality. Conclude with a Q&A session to address specific autoscaling challenges in Kubernetes environments.
Syllabus
Intro
Agenda
Recommendation life cycle
HPA
Metrics
Status
Custom Metrics
Custom Metrics in Kubernetes
State
Stabilization
Scale Controls
Golang Dock
Conditions
Backwards Compatibility
Cluster Autoscaler
Class Autoscaler Architecture
Example Scale Up
Expander
Autoscaling QA
Taught by
CNCF [Cloud Native Computing Foundation]