Intro: Autoscaling SIG
summary
Covers how to combine HPA, VPA and Cluster Autoscaler with concrete knobs: set HPA target utilization to leave a two‑minute spike buffer (example: allow 50% sudden rise by using <=65% target, or for doubling use <=50%), VPA modes are recommendation, update new pods or auto (which kills and recreates pods), and Cluster Autoscaler works with AWS, GCP, Azure, Alibaba, Baidu and OpenStack. Emphasizes keeping metrics-server, readiness/liveness probes and PodDisruptionBudgets healthy and warns VPA can shrink requests after long idle periods.
AI-written summary · about
chapters (12)
- 0:00Start
- 3:33HPA Basics
- 6:42Setting HPA target
- 7:43Best practices for HPA
- 10:11VPA Basics
- 11:35VPA Modes
- 13:16Best practices for VPA
- 16:59What's next with VPA
- 17:41CA Basics
- 19:50Supported Cloud Providers
- 20:22Best practices for CA
- 21:13What's next with CA








