Why VirtFoundry?¶
VirtFoundry turns an existing Kubernetes cluster into a multi-tenant private cloud: tenants, VPCs, security groups, VMs, volumes, snapshots, IAM, and a web UI — on top of KubeVirt, Multus, and your StorageClass (we recommend Longhorn).
Who it is for¶
- Homelab / lab teams who outgrew Proxmox and already run (or want) Kubernetes
- Platform engineers who want a CloudStack-like API/UI without operating OpenStack
- Anyone who refuses to write raw KubeVirt YAML for every tenant day-2 task
Who it is not for (yet)¶
- Single-node “install ISO and forget” with zero Kubernetes interest → stay on Proxmox
- Pure container PaaS (use your existing Kubernetes platform)
- Drop-in Proxmox Backup Server / ZFS storage admin UX
VirtFoundry vs Proxmox¶
| Dimension | Proxmox VE | VirtFoundry |
|---|---|---|
| Mental model | Hypervisor cluster + host GUI | Cloud resources: tenant, VPC, offering, API key |
| Runtime | QEMU/KVM, LXC on the host | KubeVirt VMs on Kubernetes |
| Multi-tenancy | Datacenter / ACL oriented | First-class tenants + IAM |
| Automation | API + community Terraform | First-party REST, UI, and Terraform provider |
| Storage UX | ZFS, Ceph, local disks in the product | Kubernetes StorageClass (Longhorn, …) |
| Day-2 ops | Proxmox-centric tooling | GitOps-native (Helm / Argo), Gateway API |
| Best fit | Fast bare-metal virt appliance | Private cloud control plane on an existing cluster |
Migration pitch: keep one platform (Kubernetes). Treat VMs as cloud resources, not a second silo beside the cluster.
When Proxmox still wins¶
Be honest with yourself (and adopters):
- You want an appliance ISO, not a Kubernetes learning curve
- You need PBS, ZFS datasets, or LXC as first-class product features
- Your team will not operate KubeVirt / Multus / CSI
- A handful of VMs with no multi-tenant “cloud” model is enough
VirtFoundry is the better path when you already want Kubernetes and need tenant isolation, API/UI, and GitOps around VMs.
VirtFoundry vs “just KubeVirt”¶
KubeVirt is an excellent hypervisor API. VirtFoundry is the product layer: multi-tenant isolation, networking model, catalog (templates / offerings), snapshots UX, IAM, and an opinionated Helm install.
If you only need a few VMs and are happy applying CRDs by hand, you may not need VirtFoundry.
VirtFoundry vs Harvester / similar¶
Harvester is a full HCI appliance experience. VirtFoundry assumes you bring the cluster (kubespray, kubeadm, managed Kubernetes, …) and focuses on the IaaS control plane and tenant UX.
Design principles¶
- Compose CNCF building blocks — do not reinvent the hypervisor or CSI
- API-first — UI and Terraform are clients of the same control plane
- Honest labs —
local-pathis for demos; Longhorn (or equivalent) for real disks / snapshots - Open core — Apache 2.0 core; optional enterprise services elsewhere
Try it¶
- Quickstart (under 30 min)
- Installation guide
- Minimum vs production topologies
- Traction / CNCF checklist
Also in the core repo: docs/WHY.md (keep both in sync when editing).