kube-scheduler1.35.10

kube-scheduler assigns each newly created Kubernetes pod to an appropriate node so workloads can begin running in the cluster. It continuously watches for pods that do not yet have a node assignment, evaluates eligible nodes, and records the selected placement as part of the Kubernetes control plane. Scheduling decisions consider a pod’s individual and combined resource requirements, hardware, software and policy constraints, affinity and anti-affinity rules, data locality, interference between workloads, and deadlines.

Cluster administrators and platform engineers use it when operating Kubernetes environments that must place containerized workloads according to capacity, topology, policy, and timing requirements. The scheduler is especially relevant in production clusters managed across laptops, cloud infrastructure, or data centers, including Canonical Kubernetes deployments designed for production scale.

Canonical publishes kube-scheduler as open-source software for Linux, with version 1.35.10 available for x86_64 (amd64) systems. It requires Linux with snapd and runs as part of a Kubernetes control plane, so a functioning Kubernetes environment provides its operational context. Canonical’s Kubernetes platform pairs the control plane with built-in networking, local storage, DNS, load-balancing, metrics-server, gateway, and ingress components, while Juju can automate large-scale cluster operations.

Canonical also positions this Kubernetes environment for security-maintained production use, including a built-in FIPS 140-3 mode and automated security patching through Ubuntu Pro, with support and maintenance extending up to 15 years. The scheduler’s distinguishing role is its policy-aware node selection: rather than merely finding available capacity, it balances resource, locality, compatibility, workload interaction, and deadline requirements when placing pods.