Kubernetes Scheduling Features or How Can I Make the System Do What I Want?
summary
Covers Kubernetes 1.6 scheduler mechanics with concrete knobs and limits: explains pod anti-affinity and pod affinity semantics, taints and tolerations, and node affinity now in beta; describes predicates (feasibility by resource requests only) and priorities (each returns 0–10, combined by linear weights called “waves”), shows worst‑fit vs pack‑down policies, mentions multi‑scheduler support, and outlines planned preemption/rescheduler and resource estimation. Watch if you need exact behavior, config names and predictable ways to steer placement.
AI-written summary · about
chapters (11)
- 0:00Intro
- 1:59How the scheduler works
- 13:00Notes
- 17:22Priorities
- 17:51Priority
- 20:20Spreading
- 27:54Dangers
- 30:28Summary
- 30:43State of Kubernetes
- 31:23Future Plans
- 33:18Outro
