https://shengxu.pages.dev/en/tags/kueue/Cloud architecture & DevOps notes by Shengxu: Kubernetes, Cilium, observability, LLM infra, AI agents.Hugo 0.153.2 & FixIt v0.4.0-alpha.3-20251225101113-8ffb9a95enWed, 21 Jan 2026 00:00:00 +0000https://shengxu.pages.dev/en/posts/kubernetes-1-35-native-gang-scheduling/Wed, 21 Jan 2026 00:00:00 +0000https://shengxu.pages.dev/en/posts/kubernetes-1-35-native-gang-scheduling/KubernetesAI<p>Kubernetes 1.35 introduces native Workload API and Gang Scheduling support, widely regarded as a &ldquo;kernel-level refactoring&rdquo; of cloud-native AI infrastructure. To truly grasp the significance of this upgrade, we need to look not only at what it brings but also at what it aims to replace (or merge with).</p> <p>Before v1.35, to address the &ldquo;resource deadlock&rdquo; pain point of AI training tasks, the community had actually evolved a complex &ldquo;third-party scheduler zoo.&rdquo; This article starts from the native primitives, takes stock of existing ecosystem options, and reveals the architectural evolution direction in production environments.</p>