Introduction​

Welcome to part 2 of the Talos Linux Kubernetes cluster bootstrap on the Proxmox series. Today, we will take the next step with our configuration and go through the process of enabling Cilium as our CNI (Container Network Interface) with KubeProxy replacement enabled and Cilium Hubble for network observability. We will outline basic kubectl commands to evaluate the Cilium setup alongside network tests.

We assume you already have the basic project structure from part 1 as we will extend the configuration for Cilium. To follow along, check out the part 1 post.

Introduction​

It's been a while now since I am bootstrapping RKE2 and K3s clusters on different platforms, on-prem and in the cloud, including VMware, Proxmox, Nutanix and pretty much every well-known cloud provider. This week, I have decided to take a different approach and discover something new! Bootstrap a Talos Kubernetes cluster on Proxmox using OpenTofu as the Infrastructure as Code (IaC) solution. My first interaction with Talos Linux was a couple of months back when Justin Garrison posted something about the ease of Kubernetes cluster deployment. I did not have much time back then, but here we come!

The blog post will be split into two parts. Part 1 will include a basic deployment of a Talos cluster using the out-of-box configuration, while Part 2 will contain the required configuration changes to use Cilium as our CNI. Get ready to roll up your sleeves and dive into the essentials of Talos Linux with OpenTofu on Proxmox.

Introduction​

Migration alert! For a while, I wanted to move my Docusaurus blog to a new custom domain. In the beginning, it appeared like an overwhelming task due to time constraints and of course DNS! However, after some research, I finally managed to make the migration smooth and hassle-free. This blog post is meant to save you time and frustration if you are looking to deploy a Docusaurus site on GitHub Pages with your custom domain without running into DNS problems along the way!

Funny Fact: After spending the last couple of days troubleshooting DNS issues (fun, right?), a colleague and good friend Justus surprised me with this awesome t-shirt! 😂

Now, let's dive into the configuration! 👇

Introduction​

Welcome to the the first post of the brand new Kubernetes Troubleshooting Insights section! The series of blog posts will share helpful information and troubleshooting tips for issues that might appear in a Kubernetes environment. The posts are focused on real-life scenarios from either test, staging or production environments.

In today’s blog post, we’ll explore an issue with CoreDNS setup on RKE2 clusters. Cilium CNI with Hubble were enabled for this setup. Let’s jump right in!

Introduction​

How easy is it to handle Day-2 operations with existing CI/CD tooling? Sveltos provides the ability to perform not only Day-1 operations but also helps platform administrators, tenant administrators and other operators with Day-2 operations. For example, we can use the HealthCheck and the ClusterHealthCheck features to not only watch the health of a cluster but also collect information from the managed clusters and display them in the management cluster.

In today's blog post, we will cover a way of deploying Cilium as our CNI alongside Cilium Tetragon for observability. We will then continue with a simple TracingPolicy deployment to capture socket connections and then use Sveltos to display the tracing results back to the management cluster.

The goal of the demonstration is to showcase how Sveltos can be used for different Kubernetes cluster operations based on the use case at hand.

Introduction​

In previous posts, we outlined how Sveltos allows Platform and tenant administrators to streamline Kubernetes applications and add-on deployments to a fleet of clusters. In today's blog post, we will take a step further and demonstrate how easy it is to target and update a subset of resources targeted by multiple configurations. By multiple configurations, we refer to the Sveltos ClusterProfile or Profile Custom Resource Definitions (CRDs). The demonstration focuses on day-2 operations as we provide a way to update and/or remove resources without affecting production operations.

This functionality is called tiers. Sveltos tiers provide a solution for managing the deployment priority when resources are targeted by multiple configurations. Tiers are easily integrated into existing ClusterProfile/Profile definitions alongside defining the deployment order control and straightforwardly override behaviour.

Today, we will cover the case of updating the Cilium CNI in a subnet of clusters with the label set to tier:zone2 without affecting the monitoring capabilities defined in the same ClusterProfile/Profile.

Introduction​

Sveltos is on tour! Another non-technical post describing my experience at the OSSummit Europe 2024. Apart from outlining my experience, the post will include useful resources on open-source projects I learned during the event.

About Sveltos, Gianluca Mardente and I had the chance to talk at the conference and present Sveltos and how it is used to deploy and manage different Kubernetes applications and add-ons in a Multi-Cloud setup.

In the sections below, I will outline my highlights of the conference and what I have learned, while later on, I will describe what we presented about Sveltos and where to locate the required resources.

Introduction​

Today's post will not be as technical as previous ones however, I wanted to share my experience at the Civo Navigate in Berlin. I had the chance to talk at the conference, present Sveltos and how it can be used to painlessly deploy different Kubernetes applications and monitoring capabilities on a fleet of clusters.

Apart from that, I attended many different sessions covering relevant topics (Cloud native, Security, Thought Leadership, AI) and meet fellow enthusiasts.

In the sections below, I will outline some of my highlights and afterwards provide an introduction to the Sveltos presentation alongside the useful resources.

Introduction​

In a previous post, we covered how to create an RKE2 cluster on Azure Cloud using the cloud-free credits from the Rancher UI. As this is a convenient approach to get started with Rancher, in today's post, we will demonstrate how to use OpenTofu to automate the deployment.

OpenTofu is a fork of Terraform. It is an open-source project, community-driven, and managed by the Linux Foundation. If you want to get familiar with what OpenTofu is and how to get started, check out the link here.

Additionally, we will demonstrate how easy it is to customise the Cilium configuration and enable kube-vip for LoadBalancer services from the HCL (HashiCorp Configuration Language) definition.

Introduction​

Have you ever wondered how to dynamically instantiate Kubernetes resources before deploying them to a cluster? What if I tell you there is an easy way to do it? Sveltos lets you define add-ons and applications using templates. Before deploying any resource down the managed clusters, Sveltos instantiates the templates using information gathered from the management cluster.

In a previous post, we outlined a step-by-step approach to forming a Cilium cluster mesh between two clusters. In today's post, we will demonstrate how the Sveltos templating is used to deploy a Cilium cluster mesh dynamically in one go.