gardener VS microk8s

Now, if you want to dig further on how cloud providers operate, like I said, most are actually using Kubernetes to automate the client control plane management. There is a cloud-neutral project for this out there called Gardener, they have a few architecture documents which explain the concept a bit further. In their garden metaphor, the seed cluster hosts the client control planes, and the shoot clusters are the client clusters (which are only made of worker nodes, no control-plane node). Another more specialized implementation is Kubernikus for OpenStack.

Gardener: https://github.com/gardener/gardener RH HyperShift: https://github.com/openshift/hypershift

Gardener does exactly that. One global cluster manages smaller per-region/cloud provider management clusters and those will contain the control planes of your workload clusters. This way you can have like 10 000 clusters and not deal with multi-tenant issues. One workload = 1 cluster.

We had cluster of cluster style management managing 100s of cluster. Check out https://github.com/gardener/gardener for an inspiration

How does this compare against simply using Gardener [0]?

[0] https://github.com/gardener/gardener

If you want hands-on practice you should have a running Kubernetes cluster (I used MicroK8s for this tutorial) and Helm (see how to install on Installing Helm tutorial). It is important that you understand the basics of these tools to fully understand.

And install microk8s:

MicroK8s is a lightweight, batteries included Kubernetes distribution by Canonical designed for running edge workloads which also happens to be developer-friendly and a great choice for building your own homelab. The following lab covers how to install and run MicroK8s on your own edge node running Ubuntu 22.04 LTS, deploy the NGINX web service and exposing your NGINX website to the Internet with SSL/TLS enabled using AWS resources included within the Free Tier.

One quick and easy win I can recommend, is microk8s.

I had the same problem (and there's a github issue about this: https://github.com/canonical/microk8s/issues/2186). I swapped to k3s and the usage was half of what microk8s used.

microk8s currently has a showstopping issue that makes it guaranteed to have an irrecoverable failure in HA mode. see https://github.com/canonical/microk8s/issues/3227

k0s is better but also has a lot of bugs. it's the closest to vanilla kubernetes among all the distributions.

> like the simplest GPU support

linux users should be ready to install the nvidia device plugin. if they can't do that, they're never going to succeed in running a gpu accelerated application on their cluster anyway.

> like bootstrapping

in my experience, writing all the bootstrap scripts is painful. but now that there's chatgpt, so much of the drudgery as gone away.

MicroK8S to spin up a Kubernetes cluster

They now power my microk8s/x86 cluster (in addition to my 8-node Raspberry Pi4 ARM64 microk8s cluster), microceph cluster and my LXD cluster, and all are configured with WOL, so I can bring up the cluster from any machine in the homelab, on demand.