The benefits of using managed Kubernetes, such as Google Kubernetes Engine (GKE), with contrast to self-managed Kubernetes are beyond dispute. However, it is not always feasible to leverage managed Kubernetes solutions. You might be restricted to on-premise machines because of legal constraints or have specific kubernetes add-ons (e.g. GPU sharing) that cannot be installed on top of managed Kubernetes.

At ML6, most of our clients are able to use cloud technology and thus leverage managed solutions such as GKE (Google Kubernetes Engine). Nevertheless, some of our customers do have legal constraints. In this case we can install Kubernetes on-premise in order to be able to run Machine Learning workloads in a reliable and scalable manner. It should be mentioned that hybrid-cloud products such as Google Anthos or HPE Container Platform can provide managed Kubernetes on-premise, at a cost of a licensing fee.

In this blog post, we will walk you through the process of setting up a self-managed production-grade Kubernetes cluster with Kubespray, a Kubernetes deployment tool. Kubespray is one of the mainstream tools for deploying a production-grade Kubernetes cluster.

Interested in how we set up the process of a self managed Kubernetes? Read the full interactive blogpost on our Medium blog.