A highly optimized and auto-upgradable, HA-able & Load-Balanced, Kubernetes cluster powered by k3s-on-MicroOS and deployed for peanuts on <ahref="https://hetzner.com"target="_blank">Hetzner Cloud</a> 🤑 🚀
[Hetzner Cloud](https://hetzner.com) is a good cloud provider that offers very affordable prices for cloud instances, with data center locations in both Europe and the US.
The goal of this project is to create an optimal and highly optimized Kubernetes installation that is easily maintained, secure, and automatically upgrades. We aimed for functionality as close as possible to GKE's auto-pilot.
In order to achieve this, we built it on the shoulders of giants, by choosing [openSUSE MicroOS](https://en.opensuse.org/Portal:MicroOS) as the base operating system, and [k3s](https://k3s.io/) as the Kubernetes engine.
_Please note that we are not affiliated to Hetzner, this is just an open source project striving to be an optimal solution for deploying and maintaining Kubernetes on Hetzner Cloud._
_It uses Terraform to deploy as it's easy to use, and Hetzner provides a great [Hetzner Terraform Provider](https://registry.terraform.io/providers/hetznercloud/hcloud/latest/docs)._
Then you'll need to have [terraform](https://learn.hashicorp.com/tutorials/terraform/install-cli) and [kubectl](https://kubernetes.io/docs/tasks/tools/) cli installed. The easiest way is to use the [gofish](https://gofi.sh/#install) package manager to install them.
_The Hetzner cli `hcloud` is also useful to have, mainly for debugging without having to use the Hetzner website. See how to install it [here](https://github.com/hetznercloud/cli)._
1. Create a project in your [Hetzner Cloud Console](https://console.hetzner.cloud/), and go to **Security > API Tokens** of that project to grab the API key. Take note of the key! ✅
2. Either, generate a passphrase-less ed25519 SSH key-pair for your cluster, unless you already have one that you'd like to use. Take note of the respective paths of your private and public keys. Or, for a key-pair with passphrase or a device like a Yubikey, make sure you have have an SSH agent running and your key is loaded (`ssh-add -L` to verify) and set `private_key = null` in the following step. ✅
It will take around 12 minutes to complete, and then you should see a green output with the IP addresses of the nodes. Then you can immediately kubectl into it (using the kubeconfig.yaml saved to the project's directory after the install).
_Just using the command `kubectl --kubeconfig kubeconfig.yaml` would work, but for more convenience, either create a symlink from `~/.kube/config` to `kubeconfig.yaml`, or add an export statement to your `~/.bashrc` or `~/.zshrc` file, as follows (you can get the path of kubeconfig.yaml by running `pwd`):_
⚠️ Once you start with Terraform, it's best not to change the state manually in Hetzner, otherwise when you try to scale up or down, Terraform will complain that things changed outside of it and will not be able to do it. _In the future, we will provide a tool to create bare nodes, either agents or control planes, to be joined manually._
To scale the number of nodes up or down, just make sure to properly `kubectl drain` the nodes in question first if scaling down. Then just edit these variables in `terraform.tfvars` and re-apply terraform with `terraform apply -auto-approve`.
If you want to remain HA (no downtime), it's important to **keep a number of control planes nodes of at least 3** (2 minimum to maintain quorum when 1 goes down for automated upgrades and reboot), see [Rancher's doc on HA](https://rancher.com/docs/k3s/latest/en/installation/ha-embedded/).
Otherwise, it's important to turn off automatic upgrades of the OS only (k3s can continue to update without issue) for the control-plane nodes (when 2 or less control-plane nodes), and do the maintenance yourself.
By default, MicroOS gets upgraded automatically on each node, and reboot safely via [Kured](https://github.com/weaveworks/kured) installed in the cluster.
As for k3s, it also automatically upgrades thanks to Rancher's [system upgrade controller](https://github.com/rancher/system-upgrade-controller). By default it follows the k3s `stable` channel, but you can also change to `latest` one if needed, or specify a target version to upgrade to via the upgrade plan.
You can copy and modify the [one in the templates](https://github.com/kube-hetzner/kube-hetzner/blob/master/templates/plans.yaml.tpl) for that! More on the subject in [k3s upgrades basic](https://rancher.com/docs/k3s/latest/en/upgrades/basic/).
_Also, if you had a full-blown cluster in use, it would be best to delete the whole project in your Hetzner account directly as operators or deployments may create other resources during regular operation._
This project has tried two other OS flavors before settling on MicroOS. Fedora Server, and k3OS. The latter, k3OS, is now defunct! However, our code base for it lives on in the [k3os branch](https://github.com/kube-hetzner/kube-hetzner/tree/k3os). Do not hesitate to check it out, it should still work.
There is also a branch where openSUSE MicroOS came preinstalled with the k3s RPM from devel:kubic/k3s, but we moved away from that solution as the k3s version was rarely getting updates. See the [microOS-k3s-rpm](https://github.com/kube-hetzner/kube-hetzner/tree/microOS-k3s-rpm) branch for more.