KubeFleet: Multi-Cluster Application Placement for Kubernetes
Shriira Press
A hub-and-spoke control plane for placing Kubernetes resources across a fleet of clusters, with intelligent scheduling, staged rollouts, per-cluster overrides, and drift detection.
Welcome to KubeFleet: Multi-Cluster Application Placement for Kubernetes.
KubeFleet is a CNCF sandbox project that turns a sprawl of independent Kubernetes clusters into a single, coordinated fleet. You define resources once on a lightweight hub cluster and let KubeFleet decide where they should run, propagate them to the chosen member clusters, roll changes out gradually, tailor them per cluster, and watch for anyone who drifts away from the declared state. Originally built by Microsoft as the engine behind Azure Kubernetes Fleet Manager and donated to the CNCF in early 2025, it works with any CNCF-certified distribution on cloud, on-premises, or edge. This book follows the system from the ground up: the first chapters explain what KubeFleet is and how its hub-and-spoke architecture fits together, then we register member clusters and meet the MemberCluster API. From there we dig into the ClusterResourcePlacement resource at the heart of the project, the scheduler that chooses target clusters with PickAll, PickN, and PickFixed policies, the rollout strategy that updates clusters without downtime, and the override and drift-detection machinery that keeps each cluster correct. A closing chapter gathers the practices that make a fleet pleasant to operate.
This title is part of the ShriIra library and is free to read in full, right here — our small contribution to making world-class knowledge easy to reach.
A note on reading it: open the Contents menu at the top of the reader to jump between chapters, use the Aa menu to set a comfortable text size, theme (light, sepia, or night), and single- or two-page layout. Your place is saved automatically, so you can always pick up where you left off.
We hope it serves you well.
— Shriira Press