drone-on-k8s VS kind

The accompanying code for this blog i.e. the demo sources is available on my GitHub repo. Let us clone the same on to our machine,

You can use a Kind cluster for testing. If you haven’t installed it yet, check out the Kind installation guide.

Next you will need to have access to a Kubernetes cluster but for development purposes I would advise you use a local cluster from tools like Minikube or Kind. You can visit their websites for the installation steps.

Any Kubernetes cluster 1.24+ + Kubectl configured (for this guide, I’m using an EKS. But Minikube/Kind is also welcome)

Kubernetes cluster: You need a running Kubernetes cluster that supports persistent volumes. You can use a local cluster, like kind or Minikube, or a cloud-based solution, like GKE%20orEKS or EKS. The cluster should expose ports 80 (HTTP) and 443 (HTTPS) for external access. Persistent storage should be configured to retain Keycloak data (e.g., user credentials, sessions) across restarts.

As the name suggests, Kubernetes in Docker, KinD allows you to run Kubernetes clusters locally using Docker. Each Kubernetes node is represented by a Docker container, which uses Docker’s underlying networking and storage capabilities to simulate a realistic Kubernetes setup.

We can leverage Kind’s extraPortMapping config option when creating a cluster to forward ports from the host to an ingress controller running on a node.

You should check out https://kind.sigs.k8s.io/ and https://k0sproject.io

First, make sure that you have a Kubernetes cluster up and running. If you don’t have a cluster, go ahead and install kind or minikube to get access to a local Kubernetes cluster.

Now, I need to reproduce this issue. The key point is that since these two types of Evictions behave differently, I cannot use kubectl drain or similar commands to reproduce the scenario. I need to specifically create a Node-pressure Eviction. However, I don't have a cluster to use; I do all my development on my personal computer, making it difficult to reproduce the issue. When developing Kubernetes applications locally, most people use minikube, kind, or k3d. Since I need a multi-node environment, minikube is excluded. Although it now supports multiple nodes, it's still more commonly used for single-node scenarios. Both kind and k3d use Docker containers as Kubernetes nodes. My operating system is Linux Mint, and Docker runs natively, unlike macOS where Docker runs in a virtual machine. Because the resources (memory, disk, etc.) are shared between Docker and my local machine, if I do create a Node-pressure scenario, my computer might become unusable.