Deep Dive: Autoscaling SIG
summary
Deep dive into how Vertical Pod Autoscaler and Cluster Autoscaler operate under the hood explains the VPA recommendation loop, eviction path and the CA scale up and scale down heuristics, shows the exact API objects and fields to inspect, and walks through commands and metrics to debug (kubectl describe, events, metrics-server, Prometheus traces). Demonstrations include root causes that break autoscaling, concrete fixes, and configuration knobs that change behavior such as scaleDownDelay and recommendation stabilization.
AI-written summary · about
chapters (5)
- 0:00Autoscaling SIG intro
- 10:12Community challenges and goals
- 22:50Current solutions
- 25:34New autoscaling player
- 33:27Future directions and Q&A
related talks
34mDeep Dive: SIG Autoscaling
24mSIG-Autoscaling: Introduction + HPA/VPA Updates
34mPredictive autoscaling patterns in Kubernetes
39mSIG Autoscaling Deep Dive
34mIntroduction to Autoscaling
37mDeep Dive into Autoscaling
35mAutoscaling Kubernetes Deployments: A (Mostly) Practical Guide
35mSIG Autoscaling Intro