Emerging Kubernetes tools and learning strategies

Emerging Kubernetes tools and learning strategies

Guest:

  • Daniel Bodky

This interview explores emerging Kubernetes tools, learning strategies, and future trends in cloud-native technologies.

Daniel Bodky, Platform Advocate at NETWAYS Web Services, discusses:

  • Three emerging Kubernetes tools to watch: Tetragon for security, Timoni for application deployment, and Talos as a secure Kubernetes Linux distribution.

  • The benefits of building a home lab with a bare metal Kubernetes cluster as a comprehensive learning strategy, covering aspects like networking, storage, and security.

  • Insights on Kubernetes distributions and node management, emphasizing the importance of security, centralized management, and maintaining an overview of your infrastructure.

Relevant links
Transcription

Bart: What are three Kubernetes emerging tools that you are keeping an eye on?

Daniel: I really like Cilium and Isovalent. One of the tools I'm keeping an eye on is Tetragon, as it still feels a bit new and has new features rolling in. Security gets more and more important the deeper we dive into cloud and cloud native. Another tool I'm watching is Tintor, a new way to define, distribute, and deploy applications to Kubernetes. It's written by Stefan Prodan, who is also heavily involved with Flux, and it's based on CUE lang, which could allow you to get rid of YAML in your cloud native workflow. The third tool I'd recommend is Talos, a Kubernetes Linux distribution that's incredibly small and secure by design. I use it myself, and it's great for small on-prem deployments.

Bart: One of our guests, Matthias, suggested the best way to learn Kubernetes is by doing and getting your hands dirty. He then built his own bare metal Kubernetes cluster in his spare time. What's your strategy to learn new Kubernetes tools and features?

Daniel: I'm building a home lab with a bare metal Kubernetes cluster. It's a good way to touch on all aspects of Kubernetes, considering all the things to care about and think about, just at a smaller scale. Since it's bare metal, I need to care about networking and I've chosen to use Cilium as my CNI. I also need to decide on a Kubernetes distribution to install, considering both Kubernetes and the operating system. Additionally, I need to think about securing all of that and how to get persistent storage on-prem in my home lab. I think this is a great way to learn because I get to experience all the nitty-gritty edge cases, especially since I'm running on relatively new ARM architecture, which presents some difficulties with the ARM architecture. This setup gives me the opportunity to really think about how to put things together, build things from the ground up, look at the available tooling, and maybe even contribute back to it upstream.

Bart: One of our guests, Gazal, advocated adopting Bottle Rocket OS in EKS due to its security benefits and performance improvements over traditional host OS options. What's your advice regarding Kubernetes distributions and managing nodes?

Daniel: My background is mostly in on-prem Kubernetes distributions installations, so I'm not too familiar with cloud distributions. However, I recently came across Talos, which I really like. It makes it easy to manage security issues that might arise in your Kubernetes environment. With only 12 binaries, you can manage everything using an API and a CLI tool that talks to this API, effectively locking down your whole cluster and making it very secure from the start.

If you need to run an enterprise Linux distribution on-prem, I would recommend using a central managing data plane like Rancher. This provides a good overview of the Kubernetes distributions and versions you're running, as well as the operating systems and versions. The most important thing is to maintain an overview, which enables you to catch up with security updates, make informed choices, and simplify your setup over time, making it more secure as you go.

Bart: Kubernetes is turning 10 years old this year. What should we expect in the next 10 years to come?

Daniel: I've been in the business for around two years, so I don't have a 10-year perspective. However, I think we're already seeing the ongoing commoditization of Kubernetes and the cloud, and this trend will continue. At last year's KubeCon, Internal Developer Platforms (IDPs) were a big deal, and I think we'll see their adoption rise as time goes on. People want to consume Kubernetes without having to maintain, manage, and build it from scratch every time they use it. As we saw in the keynote earlier today, AI and ML are here to stay, and we're living in the age of AI. Kubernetes can be a catalyst for better and faster training of Large Language Models (LLMs), for example, which I think will be a huge advantage going forward.

Bart: What's next for you?

Daniel: Well, I just switched into my first DevRel position. I'm a platform advocate now, so I'm still figuring things out and trying to talk to many people in the space, including thought leaders and people who are managing DevRel, to hit the ground running more effectively as I go.

Podcast episodes mentioned in this interview