Back to Vcluster

README

README.md

0.35.120.0 KB
Original Source
<div align="center"> <a href="https://www.vcluster.com"> <picture> <source media="(prefers-color-scheme: dark)" srcset="docs/static/media/vcluster_horizontal_orange_white.svg"> <source media="(prefers-color-scheme: light)" srcset="docs/static/media/vcluster_horizontal_orange_black.svg">
</picture>
</a> <p><strong>Tenant Clusters for Production Kubernetes and AI Infrastructure</strong></p> <p><em>Virtual control planes, real isolation โ€” from a single node to 100K-GPU superclusters.</em></p>

Website โ€ข Quickstart โ€ข Documentation โ€ข Blog โ€ข Slack

<a href="https://www.cncf.io/training/certification/software-conformance/"></a> ย ย ย ย  <a href="https://github.com/cncf/k8s-ai-conformance/tree/main/v1.35/vcluster-private-nodes"></a>

CNCF Certified Kubernetes โ€” Distribution ยท Kubernetes AI Conformant

</div>

What is vCluster?

vCluster creates Tenant Clusters โ€” fully isolated Kubernetes environments that run on top of a Control Plane Cluster or standalone on dedicated infrastructure or bare metal. Each tenant gets its own API server, CRDs, and RBAC, with a cluster experience indistinguishable from a dedicated Kubernetes cluster.

Built for production. Trusted in production. 40M+ Tenant Clusters deployed by teams at Adobe, CoreWeave, NVIDIA, Lintasarta, Atlan, Deloitte, and hundreds of AI clouds, AI factories, and Fortune 500 platform organizations.

CNCF Certified Kubernetes โ€” Distribution and Kubernetes AI Conformant โ€” every Tenant Cluster is upstream Kubernetes with no vendor lockโ€‘in, validated for portable AI/ML workloads (training, inference, agentic).

The public-cloud experience, on your own infrastructure. Give every team the Kubernetes they need โ€” with strict isolation, hardware-aware scheduling, and zero tenant sprawl โ€” whether you run one region or 100K GPUs.

<div align="center">

</div>

๐Ÿš€ Quick Start

bash
# Install vCluster CLI
brew install loft-sh/tap/vcluster

# Create a Tenant Cluster
vcluster create my-vcluster --namespace team-x

# Use kubectl as usual โ€” you're now in your Tenant Cluster
kubectl get namespaces

Prerequisites: A running Kubernetes cluster and kubectl configured.

๐Ÿ‘‰ Full Quickstart Guide

๐Ÿณ Run Locally with Docker โ€” vind

No Kubernetes cluster? Run vCluster directly on Docker with vind (vCluster in Docker) โ€” like kind, but with the full vCluster feature set (UI, sleep/resume, LoadBalancer, image cache):

bash
vcluster create my-vcluster --driver docker
kubectl get namespaces

๐ŸŽฎ Try in the Browser

๐ŸŽ vCluster Free Tier

Real usage, not a gated demo. Unlimited Tenant Clusters up to 64 CPUs / 32 GPUs, plus the full vCluster Platform UI โ€” for free. Get Started Free โ†’


๐Ÿ†• What's New

VersionFeatureDescription
v0.34Multi-Region Platform & Standalone SnapshotsActive/active vCluster Platform across regions (Route 53 + RDS), Standalone snapshots (S3 / OCI / local), first-class template parameters
v0.33Enterprise Reliability & StorageAutomatic leaf-cert regeneration, Azure Blob snapshot destinations, workload-level sleep annotations
v0.32Docker Driver & DRARun vCluster on Docker, Dynamic Resource Allocation (DRA) for GPU workloads, in-place pod resizing
v0.31Snapshots & Cross-Cluster APIsExpanded snapshot/restore lifecycle, PDBs for Tenant Cluster control planes, cross-cluster resource proxying
v0.30vCluster VPN & Netris IntegrationTailscale-powered overlay networking and automated hardware isolation via Netris
v0.27โ€“v0.29Architecture FoundationsPrivate Nodes (v0.27, CNI/CSI isolation), Auto Nodes (v0.28, Karpenter autoscaling), Standalone Mode (v0.29, no Control Plane Cluster โ€” dedicated infrastructure or bare metal)

๐Ÿ‘‰ Full Changelog


๐ŸŽฏ Use Cases

Use CaseDescriptionLearn More
AI FactoryRun AI on-prem where your data and GPUs live. Give every team the GPU access they need without multiplying infrastructure.View โ†’
AI Cloud ProvidersLaunch a hyperscaler-like Kubernetes experience for your GPU customers. Isolated, production-grade, in minutes.View โ†’
Internal GPU PlatformMaximize GPU utilization without sacrificing isolation. Self-service Kubernetes for AI/ML teams.View โ†’
Bare Metal KubernetesRun production Kubernetes on bare metal with zero VMs. Isolation without expensive virtualization overhead.View โ†’
Software VendorsShip Kubernetes-native products. Each customer gets their own isolated Tenant Cluster.View โ†’
Environments & Cost SavingsConsolidate clusters, pause idle workloads with sleep mode, and cut Kubernetes cost at scale.View โ†’

๐Ÿ—๏ธ Architectures

vCluster supports multiple deployment architectures. Each builds on the previous, offering progressively stronger isolation โ€” from dense shared infrastructure to fully standalone deployments on dedicated infrastructure or bare metal.

Architecture Comparison

Shared NodesDedicated NodesPrivate NodesStandalone
Control Plane ClusterRequiredRequiredRequiredNot Required
Node IsolationโŒโœ…โœ…โœ…
CNI/CSI IsolationโŒโŒโœ…โœ…
Bare Metal Readyโ€”โ€”โœ…โœ…
Best ForDev/test, densityProduction tenantsCompliance, GPUAI factories, edge

๐Ÿ‘‰ Full Architecture Guide

Minimal Configuration

<details> <summary>๐Ÿ”น Shared Nodes โ€” Maximum density, minimum cost</summary> Tenant Clusters share the Control Plane Cluster's nodes. Workloads run as regular pods in a namespace. <div align="center"> </div>
yaml
sync:
  fromHost:
    nodes:
      enabled: false  # Uses pseudo nodes
</details> <details> <summary>๐Ÿ”น Dedicated Nodes โ€” Isolated compute on labeled node pools</summary> Tenant Clusters get their own set of labeled nodes on the Control Plane Cluster. Workloads are isolated but still managed by the Control Plane Cluster. <div align="center"> </div>
yaml
sync:
  fromHost:
    nodes:
      enabled: true
      selector:
        labels:
          tenant: my-tenant
</details> <details> <summary>๐Ÿ”น Private Nodes <sup>v0.27+</sup> โ€” Full CNI/CSI isolation</summary> External nodes join the Tenant Cluster directly with their own CNI, CSI, and networking stack. Complete workload isolation from the Control Plane Cluster. <div align="center"> </div>
yaml
privateNodes:
  enabled: true
controlPlane:
  service:
    spec:
      type: NodePort
</details> <details> <summary>๐Ÿ”น vCluster Standalone <sup>v0.29+</sup> โ€” No Control Plane Cluster required</summary> Run vCluster without any Control Plane Cluster. Deploy the Virtual Control Plane directly on bare metal or VMs. The highest level of isolation โ€” vCluster becomes the cluster. <div align="center"> </div>
yaml
controlPlane:
  standalone:
    enabled: true
    joinNode:
      enabled: true
privateNodes:
  enabled: true
</details> <details> <summary>โšก Auto Nodes <sup>v0.28+</sup> โ€” Karpenter-powered dynamic autoscaling</summary> Automatically provision and deprovision private nodes based on workload demand. Works across public cloud, private cloud, hybrid, and bare metal environments. <div align="center"> </div>
yaml
autoNodes:
  enabled: true
  nodeProvider: <provider>
privateNodes:
  enabled: true
</details>

โœจ Key Features

FeatureDescription
๐ŸŽ›๏ธ Isolated Virtual Control PlaneEach Tenant Cluster gets its own API server, controller manager, and data store โ€” complete Kubernetes API isolation
๐Ÿ”— Shared Platform Stack (Shared / Dedicated Nodes)Leverage the Control Plane Cluster's CNI, CSI, ingress, and other infrastructure โ€” no duplicate platform components
๐Ÿ”’ Strong Tenant IsolationTenants get admin access inside their Tenant Cluster while having minimal permissions on the Control Plane Cluster
๐Ÿ”„ Resource Syncing (Shared / Dedicated Nodes)Bidirectional sync of any Kubernetes resource โ€” pods, services, secrets, configmaps, CRDs, and more
๐Ÿ’ค Sleep ModePause inactive Tenant Clusters to save resources. Instant wake when needed
๐Ÿ–ฅ๏ธ Standalone DeploymentRun without a Control Plane Cluster on dedicated infrastructure or bare metal โ€” purpose-built for AI factories and on-prem GPU fleets
๐Ÿงฉ IntegrationsNative support for cert-manager, external-secrets, KubeVirt, Istio, and metrics-server (host-side integrations apply in Shared / Dedicated Nodes modes)
๐Ÿ“Š High AvailabilityMultiple replicas with leader election. Embedded etcd or external databases (PostgreSQL, MySQL, RDS)

Shared Platform Stack, Resource Syncing, and host-cluster integrations apply only in Shared and Dedicated Nodes modes, where the Tenant Cluster shares the Control Plane Cluster's CNI, CSI, and platform stack. Private Nodes and Standalone deployments bring their own CNI, CSI, and platform components.


๐ŸŒ The vCluster Platform

vCluster is the foundation of a broader platform for running production Kubernetes and AI infrastructure on your own hardware โ€” from a single rack to 100K-GPU supercomputers.

ProductWhat it does
vClusterTenant Clusters โ€” Virtual Control Planes with API, data, and (optionally) network isolation
vNodeRuntime-level isolation. Kernel-enforced boundaries (seccomp, cgroups, namespaces, AppArmor) without VM overhead
vMetalZero-touch bare metal provisioning for GPU fleets. Turns GPU racks into a cloud platform
Netris (integration)Hardware-enforced network isolation via programmatic VLANs, VRFs, and ACLs

Together these provide a complete foundation for AI factories โ€” certified Kubernetes stacks, isolated Tenant Clusters, runtime workload sandboxing, and GPU infrastructure operations โ€” the same pattern used to run production AI on hundreds of AI clouds and Fortune 500 on-prem platforms.


๐Ÿข Trusted By

<table> <tr> <td align="center"><a href="https://www.vcluster.com/case-studies/atlan"><strong>Atlan</strong></a> 100 โ†’ 1 clusters</td> <td align="center"><a href="https://www.vcluster.com/case-studies/aussie-broadband"><strong>Aussie Broadband</strong></a> 99% faster provisioning</td> <td align="center"><a href="https://www.vcluster.com/case-studies/coreweave"><strong>CoreWeave</strong></a> GPU cloud at scale</td> </tr> <tr> <td align="center"><a href="https://www.vcluster.com/case-studies/lintasarta"><strong>Lintasarta</strong></a> 170+ Tenant Clusters in prod</td> <td align="center"><a href="https://www.vcluster.com/case-studies/fortune-500-insurance-company"><strong>Fortune 500 Insurance</strong></a> 70% reduction in Kubernetes cost</td> <td align="center"><a href="https://www.vcluster.com/case-studies/scanmetrix"><strong>Scanmetrix</strong></a> 99% faster deployments</td> </tr> <tr> <td align="center"><a href="https://www.vcluster.com/case-studies/deloitte"><strong>Deloitte</strong></a> Enterprise K8s platform</td> <td align="center"><a href="https://www.vcluster.com/case-studies/ada-cx"><strong>Ada</strong></a> 10x Developer Productivity</td> <td align="center"><a href="https://www.vcluster.com/case-studies/trade-connectors"><strong>Trade Connectors</strong></a> 50% reduction in K8s ops cost</td> </tr> </table>

Also used by: NVIDIA, ABBYY, Precisely, Shipwire, and many more โ€” with 50+ GPU clouds and Fortune 500s running vCluster in production.

๐Ÿ‘‰ View All Case Studies


๐Ÿ“š Learn More

<details> <summary><strong>๐ŸŽค Conference Talks</strong></summary>
EventSpeakerTitleLink
KubeCon NA 2025 (Keynote)Lukas GenteleAutoscaling GPU Clusters Anywhere โ€” Hyperscalers, Neoclouds & BaremetalWatch
Platform Engineering Day NA 2025 (Keynote)Saiyam PathakAI-Ready Platforms: Scaling Teams Without Scaling CostsWatch
Rejekts NA 2025Hrittik Roy, Saiyam PathakBeyond the Default Scheduler: Navigating GPU MultiTenancy in AI EraWatch
KubeCon EU 2025Paco Xu, Saiyam PathakA Huge Cluster or Multi-Clusters? Identifying the BottleneckWatch
HashiConf 2025Scott McAllisterGPU sharing done right: Secrets, security, and scaling with Vault and vClusterWatch
FOSDEM 2025Hrittik Roy, Saiyam PathakAccelerating CI Pipelines: Rapid Kubernetes Testing with vClusterWatch
KubeCon India 2024 (Keynote)Saiyam PathakFrom Outage To Observability: Lessons From a Kubernetes MeltdownWatch
CNCF Book Club 2024Marc BoorshteinKubernetes - An Enterprise Guide (vCluster)Watch
KCD NYC 2024Lukas GenteleTenant Autonomy & Isolation In Multi-Tenant Kubernetes ClustersWatch
KubeCon EU 2023Ilia Medvedev, Kostis KapelonisHow We Securely Scaled Multi-Tenancy with VCluster, Crossplane, and Argo CDWatch
KubeCon NA 2022Joseph Sandoval, Dan GarfieldHow Adobe Planned For Scale With Argo CD, Cluster API, And VClusterWatch
KubeCon NA 2022Whitney Lee, Mauricio SalatinoWhat a RUSH! Let's Deploy Straight to Production!Watch
TGI Kubernetes 2022TGITGI Kubernetes 188: vClusterWatch
Mirantis Tech Talks 2022MirantisMulti-tenancy & Isolation using Virtual Clusters (vCluster) in K8sWatch
Solo Webinar 2022Rich Burroughs, Fabian KellerSpeed your Istio development environment with vClusterWatch
KubeCon NA 2021Lukas GenteleBeyond Namespaces: Virtual Clusters are the Future of Multi-TenancyWatch
</details> <details> <summary><strong>๐ŸŽฌ Community Voice</strong></summary>
ChannelSpeakerTitleLink
TeKanAid 2024TeKanAidGetting Started with vCluster: Build Your IDP with Backstage, Crossplane, and ArgoCDWatch
Rawkode 2021David McKay, Lukas GenteleHands on Introduction to vClusterWatch
Kubesimplify 2021Saiyam Pathak, Lukas GenteleLet's Learn vClusterWatch
TechWorld with Nana 2021NanaBuild your Self-Service Kubernetes Platform with Virtual ClustersWatch
DevOps Toolkit 2021Viktor FarcicHow To Create Virtual Kubernetes ClustersWatch
</details>

๐Ÿ‘‰ YouTube Channel โ€ข Blog


๐Ÿค Contributing

We welcome contributions! Check out our Contributing Guide to get started.


๐Ÿค Connect with us


๐Ÿ“œ License

vCluster is licensed under the Apache 2.0 License.


<div align="center">

ยฉ 2026 Loft Labs. All rights reserved.

Made with โค๏ธ by the vCluster community.

โญ Star us on GitHub โ€” it helps!

</div>