Skip to content
DCS

EXPLORE DCS

K3s clusters for learning Kubernetes

K3s is a lightweight Kubernetes distribution. DCS brings cluster creation, connection details and HTTP publishing into the console so you can work with container applications and learn the Kubernetes workflow.

A place for containers, manifests and Helm charts

Choose K3s when you want to organise an application as Kubernetes workloads and Services, practise deployments or try a Helm chart. Choose a VM instead if your main task is to install software directly into an operating system.

The selected cluster plan defines the node count and resource sizes. Automatic K3s installation prepares the cluster, but you still manage your workloads, application permissions, images and data. Multiple nodes do not amount to a high-availability promise.

Connect over the private network

The cluster’s Connect dialog separates device connectivity from Kubernetes credentials. Enrol your computer into the private network, then download your personal kubeconfig. Downloading the kubeconfig alone does not establish a network connection.

The kubeconfig gives administrator access to that cluster and has an expiry. Keep it private, reopen Connect when you need replacement credentials, and revoke access when it is no longer required. The browser terminal offers another way to work when the control-plane node and your permissions allow it.

The deployment workflow

  1. When customer deployments open, select a cluster plan covered by your organisation’s free-beta grant.
  2. Choose the project and private network, review the allocation and follow the order until the cluster is ready.
  3. Open Connect and follow the instructions for your computer. Check node readiness with kubectl before applying a workload.
  4. Deploy your application using Kubernetes manifests or Helm. Publish its HTTP Service separately if you want a public HTTPS address.

Understand the beta limits

New customer deployments are currently paused while beta launch checks are completed. The free beta requires an invitation and no payment details. Host capacity and organisation quotas apply independently of grants.

Use noncritical workloads. This is not a fully managed production Kubernetes or disaster-recovery offering. Deleting a cluster removes its nodes and the data stored on them; retain your own copies before cleanup.