Engineer Nelson Elhage offers several reasons Kubernetes is so complex but this does at least mean that multiple companies offer tools to try to help you master it.
The Google-backed container-management system is famously difficult, even to spell or pronounce. (It’s often called “k8s” for short: since “kubernetes” is 10 letters long, “k8s” signifies “k” + eight letters + “s”, and is pronounced “kates”.) The Mountain View mammoth even commissioned a comic to explain what it is. (It’s long, but quite good.)
K8s is a set of tools for managing clusters – but not everyone has a spare cluster lying around that they can play with.
How many K8s would you like?
One option is to build a virtual cluster on a local PC.
The Kubernetes project offers minikube for this, but it’s not the only option.
SUSE subsidiary Rancher offers K3s, a CNCF sandbox project which compiles a set of components into a single binary for easier installation. (“K3s” – or so says Rancher – is meant to denote that it’s half the size of “k8s”, being a five-letter word stylized as K3s rather than the 10 letters abbreviated by the stylisation k8s, and officially there isn’t a pronunciation.)
If you want a GUI thrown in as well, it just launched Rancher Desktop.
Mirantis, which bought Docker’s enterprise division a few years ago, offers K0s (“kay-zeros”), the successor to its older Pharos Kubernetes. Like K3s, it’s a single binary, but it’s about three times larger at ~150MB to K3s’~50MB, as it includes functionality that K3s banishes to add-on modules.
Yo dawg, I heard you like virtualization
Turning the concept inside out, if you’re already using Docker and Go and have a host machine to hand, there’s also kind, which runs K8s nodes inside Docker containers.
Perhaps you’re already running multiple K8s clusters and have the infrastructure ready, but you don’t have admin-level access. Loft Labs has you covered: the newly announced vcluster lets you run clusters inside your cluster, so you can have root on your own private virtual K8s setup.
Whether local on Arm, x86, Linux, Mac or Windows, on Docker, or in local VMs, or remote VMs, or a remote K8s cluster, someone somewhere has an allegedly easy-to-use offering to help you dip your toe in the water. After all, it’s a brave company that says it doesn’t need Kubernetes. ®