Roadmap

First release

• ClusterClasses for kubeadm, RKE2 and k3s, each passing the end-to-end suite.
• In every cluster: Proxmox CSI plugin, Proxmox cloud controller manager, control-plane virtual IP, CNI (the distribution's own or one chosen per cluster).
• Add-ons: Rancher, with Kubehenge's clusters registered in it, and Headlamp.
• The VM templates for every distribution (kubehenge template).
• The appliance: one line in the Proxmox VE shell installs it as an LXC container, whose interface creates the management cluster and keeps it upgraded; the same management cluster from one Terraform/OpenTofu apply for those who prefer it.
• The interface: create, list, scale, upgrade and delete clusters, choose the distribution and add-ons, download a kubeconfig, with permissions following Proxmox VE's on each cluster's pool.
• Upgrades without a shell: the appliance upgrades itself from a button, or automatically to new patch or minor releases of Kubehenge, installing only signed releases; each cluster can upgrade Kubernetes automatically inside a maintenance window its owner sets.
• The optional Kubernetes entry inside Proxmox VE's own interface.
• Placement: a cluster's VMs spread over every online node, or over the nodes chosen by hand or from one of Proxmox VE's HA node-affinity rules; the cluster page shows which node each VM runs on and warns when control-plane VMs share one; optionally, Proxmox VE HA keeps the control-plane VMs apart and restarts them after a node failure.

Next

• A Talos ClusterClass, once Cluster API's Talos providers support ClusterClass in a release.
• Add-ons: CIS hardening per distribution, pod security through Pod Security Admission or Kyverno, monitoring.

Later

Driven by what users ask for. Candidates: clusters across several Proxmox VE nodes and clusters, backup and restore with Proxmox Backup Server, more add-ons, and OKD clusters once Cluster API's OpenShift Assisted providers support ClusterClass and work with the Proxmox VE provider.