SIG-Node: Intro and Deep Dive
Talkat KubeCon + CloudNativeCon Europe 2025 · on Wednesday, 2025-04-02· at 15:15· Level 3 | ICC Capital Suite 14-16·
summary
SIG-Node reports concrete progress: in-place pod resize beta ships (handles allocatable checks and blocks resizes when node reserved resources are insufficient), structured DRA parameters beta lands for device allocation while DRA stays device-focused and device overprovisioning remains unsupported, and SIG-Node graduates many KEPs—about 24 KEPs in the 1.33 cycle. PSI metrics reach alpha, CPU manager LLC-awareness benchmarks show 20–30% gains on specific workloads, and device plugin vs DRA semantics explain crash-loop allocation behavior when devices are slow to provision.
AI-written summary · about
chapters (10)
- 0:00SIG-Node intro and team
- 1:05SIG-Node in Kubernetes
- 3:31Kubelet resource management
- 8:51Pod resource allocation
- 12:02Resource challenges today
- 18:47Resource expressiveness trends
- 21:57Open questions and future
- 23:03SIG-Node achievements
- 30:46Contributing to SIG-Node
- 33:37Q&A: confidential containers
related talks
28mKubernetes SIG Node Intro And Deep Dive
41mSIG Node Deep Dive
34mKubernetes SIG Node Intro and Deep Dive
31mKubenetes SIG Node Intro and Deep Dive
20mKubernetes SIG Node Intro and Deep Dive
19mKubernetes SIG Storage Intro & Deep Dive
43mKubernetes SIG Node Intro and Deep Dive
42mGuide Us on Next Steps in Kubernetes Device Manage...
