hetzner-k3s v1.0.1 is out with autoscaling support! (Kubernetes on Hetzner Cloud)

I'm so happy! Many people requested autoscaling as a feature and I released a new version today which adds this. 🎉

The repo is at https://github.com/vitobotta/hetzner-k3s for those of you who use it or are interested.

For those not familiar with it, it's my CLI tool to super quickly create and manage cheap Kubernetes clusters in Hetzner Cloud.

With this tool you can create a fully functional clusters in just a few minutes, much faster than with Terraform/Ansible or similar tools and without their complexity. You only need to download a binary and write a super simple YAML config file and that's it.

The tool:

  • creates all the infrastructure resources required (servers, load balancers, networks, firewalls)
  • deploys K3s, which is a lightweight, fully conformant Kubernetes distro
  • deploys drivers to provision load balancers and persistent volumes out of the box
  • deploys a system upgrade controller which makes upgrading to a new version of Kubernetes ridiculously easy and fast with no apps downtime
  • deploys the cluster autoscaler, required for autoscaling
  • supports highly available control planes with multiple masters
  • supports multi-region node pools for higher availability

All of this in a few minutes even for production grade clusters.

With the most recent changes, with this tool you can create a cluster that behaves like a managed Kubernetes service with autoscaling, HA and multi-region, for a fraction of the cost.

If you work with Kubernetes and are looking to create cheap clusters easily, give it a try and lemme know what you think :)

Comments

  • This is fantastic, awesome work here! Might try this out.

  • @fluttershy said:
    This is fantastic, awesome work here! Might try this out.

    Thanks! Looking forward to any feedback you may have 👍

  • Approximately, how much will it cost me to run a small cluster?

  • @AaronSS said:
    Approximately, how much will it cost me to run a small cluster?

    It depends on what you are going to run on it and how much cpu and memory your workloads need. Do you have an idea about that?

Sign In or Register to comment.