Kubernetes as a service on Superphenix
Superphenix is designed to offer Kubernetes as a Service (KaaS) so tenants can run managed Kubernetes clusters on the same virtualization, storage, and network stack as your IaaS workloads.
Main features
VM pools (worker nodes)
Tenant clusters are backed by pools of VMs where you define:
- VM sizing: CPU, memory, and disk for control-plane and worker machines.
- Networking: Node pools are attached to your VPCs and subnets; pods and services use the integrated CNI on top of that underlay.
Platform integration
| Capability | Behavior |
|---|---|
| CNI | Integrated CNI: cluster networking is provided by the platform stack (same SDN as VMs where applicable). |
| CSI | Integrated CSI: persistent volumes map to the platform storage layer. |
| Live migration | Worker VMs can be live-migrated like any other instance when policy allows, for node maintenance without hard downtime for stateless workloads. |
| Disaster recovery | Controlplane data and worker nodes can be backed up and replicated allowing for quick recovery in case of outages or catastrophic failure. |
Operations
- Create, scale, upgrade, and delete clusters via GitOps and console.
- Use the same organization / project RBAC as the rest of the platform; see Organizations and projects.
Foundation used
| Foundation | How Kubernetes uses it |
|---|---|
| Virtualization | VM pools for nodes (Compute) |
| Network | VPCs, subnets, NAT, LB, firewall (Network) |
| Storage | CSI-backed volumes (Storage) |
Version support
Each version of Superphenix will support 3 minor Kubernetes versions at a time. You may encounter different supported versions for a single Superphenix deployment as some AZs might be more up to date than others.