Managed Kubernetes on Hikube

Hikube offers a managed Kubernetes service designed to provide a highly available, secure and high-performance infrastructure. The control plane is fully managed by the platform, while the worker nodes are deployed in your project as virtual machines.

Clusters are created, modified and deleted from the Hikube console, menu Infrastructure > Kubernetes. Once the cluster is ready, a project administrator downloads its kubeconfig from the console, and you work in the cluster with your usual tools (kubectl, helm, SDK client, etc.).

---

Architecture

Hikube Kubernetes clusters rely on a multi-datacenter infrastructure (3 Swiss sites) that ensures replication, fault tolerance and service continuity.

• Control Plane: hosted and operated by Hikube. It consists of:
  • kube-apiserver
  • etcd
  • kube-scheduler
  • kube-controller-manager
• Worker nodes: virtual machines in your project, grouped into node groups
• Networking: Cilium CNI, support for LoadBalancer Services, Ingress and NetworkPolicy
• Storage: persistent volumes replicated across the 3 datacenters
• Addons: Cert-Manager, Ingress NGINX, Flux CD, monitoring agents, Velero, GPU Operator, etc.
• Kubernetes versions: you choose the version among those offered by the platform

---

What you configure in the console

The Create cluster wizard groups the configuration into four steps:

Step	What you define
General	Cluster name, Kubernetes version, API endpoint (optional), size and number of control plane instances
Nodes	One or more node groups: name, instance type, node disk size, minimum and maximum number of nodes, exposure on the internet, GPU
Addons	Enabling the cluster addons and optionally overriding their Helm values
Summary	Summary before deployment

Each field is described in detail in the concepts and the quick start.

---

How it works in detail

Control Plane

• Managed by Hikube, with no maintenance required on your side
• Sized by an instance size (Control Plane Instance Size: s1.small or s1.medium, applied to its four components) and a number of instances (Control Plane High Availability: 1, 3 or 5)
• Access through the standard Kubernetes API (kubectl, SDK client, etc.) with the kubeconfig downloaded from the console

Node groups

Node groups let you adapt resources to your workloads. Each group has its own instance type and its own autoscaling bounds.

• Autoscaling: minimum and maximum number of nodes per group
• GPU support: NVIDIA GPUs attached to the nodes of a group, chosen in the wizard
• Instance types: Standard (S), Universal (U) and Memory (M) series

---

Persistent storage

Persistent volumes (PVCs) created in the cluster use the replicated storage class:

• Automatic replication across the 3 Swiss datacenters
• Dynamic provisioning of persistent volumes
• Native fault tolerance and high availability

Example PVC to deploy in your cluster:

pvc.yaml

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: my-data
spec:
  accessModes:
    - ReadWriteOnce
  storageClassName: replicated
  resources:
    requests:
      storage: 20Gi

---

Kubernetes versions

• The version is chosen when the cluster is created, among those offered by the platform (the most recent one is preselected)
• Upgrades are done from the cluster's edit page (see How to upgrade a cluster)

---

Built-in addons

Cert-Manager

• Automated management of SSL/TLS certificates
• Support for Let's Encrypt and private authorities
• Automatic renewal

Ingress NGINX

• Built-in ingress controller, exposed by a LoadBalancer Service
• Deployed on the node groups exposed on the internet

Flux CD (GitOps)

• Continuous synchronization with your Git repositories
• Automated deployment and rollback

Monitoring Agents

• Collection of the cluster's metrics and logs (VictoriaMetrics Agent, Fluent Bit, kube-state-metrics, node exporter)

The full list is in the Plugins section.

---

Example use cases

Use case	Recommended node group
Web applications	Standard series (S), 2 to 10 nodes, group exposed on the internet to host the Ingress
ML/AI workloads	Universal series (U) with GPU, GPU Operator addon enabled
Critical applications	At least 3 nodes minimum, highly available control plane (3 instances)

---

Resources

• Concepts and architecture: understand how a Hikube Kubernetes cluster is deployed
• Quick start: create your first cluster from the console

---

Key points

• Managed control plane: no master maintenance required
• Nodes in your project: full control over the workers
• Autoscaling: dynamic adjustment to the load
• Multi-datacenter: native high availability and replication
• Full compatibility: standard Kubernetes API