Production-grade Kubernetes

The open-source orchestration system for cloud native infrastructures.

The easiest way to automate the deployment, scaling, lifecycle and management of your applications. Operate the latest Kubernetes, from the experts behind Ubuntu and the Kubernetes community.

Works anywhere you run Ubuntu

  • AWS
  • Azure
  • Google Cloud Platform
  • vmware
  • lxd
  • openstack
  • joyent
  • maas

Best practices across core components

Latest upstream Kubernetes

  • TLS encryption between machines and clients
  • High availability and load balanced control plane
  • In-place upgrades for Kubernetes
  • Dynamic scaling of masters and nodes
  • Choice of SDN: flannel, calico, weave


  • Distributed key value store
  • In-place upgrades
  • Dynamic scaling
  • TLS encryption


  • Certificate authority
  • Serves signed certificates to the requesting units of the cluster

Portable, validated, scalable

  • Contains all the operational tooling needed to manage a production cluster — including upgrades and elastic scaling
  • Upgrade, pause and resume, snapshot and restore your nodes
  • Extensible and self-healing system
  • Tested on a wide range of infrastructures, so you can avoid cloud lock-in and federate workloads between clouds or on-premise
  • Optimised portability across all environments
  • 100% validated against the Kubernetes Conformance Tests

1 Pick your cluster from production-grade blueprints

Staging cluster ›

Start with a Kubernetes cluster of two machines — one master node and one worker node.

Production cluster ›

Progress to a highly available Kubernetes cluster — 2 masters, 3 workers, 3 etc nodes and API load balancer.

Deep Learning cluster

Accelerate Deep Learning workloads by adding GPUs at scale, with totally automated deployment.

Start GPU tutorial

Or try it on a single machine

Get Kubernetes up and running on a single system with LXD, or a public cloud such as AWS, GCE or Azure.

2 Add extra components

Prometheus ›

Monitor, profile and debug your applications.

ElasticSearch ›

Centralize all logs and metrics and get awesome dashboards for your cluster.

Ceph ›

Run stateful applications on bare metal or private clouds.

3 Put your cluster to work

The easy way to commoditise GPUs for Kubernetes

Setting up your Deep Learning framework is now easier than ever with the new GPU integration of the Canonical Distribution of Kubernetes. Perfect for production grade, on bare metal or in the cloud.

Build a transcoding platform in minutes

The Canonical Distribution of Kubernetes enables automatic discovery of GPU devices. Work as well with LXD to get a fine grain control mechanism to orchestrate video workflows and maximize the throughput of your clusters.

Transform your solution into a private PaaS

Kubernetes can be complemented by another layer to transform it into a PaaS, to maximise the proficiency of the development teams. It will provide complete freedom to deploy and manage your applications.

Support and customisation services


From $600 / per node per year
Enterprise support for Kubernetes is provided by Canonical in partnership with Google where customers gain access to a global pool of knowledge and expertise.

BootStack managed Kubernetes

$12 / per node per day
Request a fully managed Kubernetes: we provide the in-house expertise to stand-up and scale your cloud. And, when you're ready, we will hand over operational control.

Get in touch

To contact us for more information or to discuss your specific needs, email us at