Blogs

Below you can view a complete list of my uploaded blogs. I also have a little table showing an overview of what my cluster consists of.

For more information see resource planning.

Node / Type	RAM Allocatable	Pod Capacity	Example Workloads
16GB / CP1	~14GB	50-80 pods	Etcd, Prometheus
16GB / CP2	~14GB	50-80 pods	Same as CP1
16GB / CP3	~14GB	50-80 pods	Same as CP1
8GB / Worker 1 (arm)	6GB	20-40 pods	Pi-hole, Home Assistant monitoring agents, lightweight apps
40GB / Worker 2 (x86)	36GB	100-200 pods	StatefulSets, Postgresql, heavy services

Kubernetes Releases v1.35

Kubernetes Releases version 1.35 On 17 December 2025 Kubernetes released version 1.35 code named “Timbernetes”, which is presented as a turning-point release that modernizes cluster operations, tightens security, and brings long-awaited vertical scaling and AI-focused scheduling into the core platform. Timbernetes has a theme of deepening the platform’s security posture while simplifying the operational model for platform teams. Release theme and scope Timbernetes uses a “world tree” metaphor to emphasize maturity, deep roots in stability/security, and new “branches” for AI and edge workloads. ...

Control Plane Setup

This blog walks though how I went about setting up kubeadm on my primary control plane node. This concerns setting up the control plane nodes which run a fresh installation of Pi OS. I’ve already reserved IPs for all Raspberry Pis on my router. For my homelab static IPs will ensure stable Kubernetes node communication, ExternalDNS updates, and Harbor registry access without lease expiration disruptions. If you want to see the polished end result, the install script is located here. ...

Resource Planning

Resource Planning My five node cluster currently consists of: Three control plane nodes (cp1, cp2 and cp3: 16GB ram Raspberry Pi 5s): all running Raspberry Pi OS (Debian 13) in headless mode so are lightweight and Pi optimised. Etcd requires the most RAM (3GB for small clusters like mine with <100 pods), while other components should typically stay under 1GB each (with Prometheus at roughly 3gb as an exception) with my multi-node control plane set up which distributes load, so 16GB handles 20-50 pods across nodes without OOM issues. Two worker nodes (wn1, wn2), one runs Debian and the other Ubuntu. I have mixed 8GB and 36GB worker nodes which work fine for my Kubernetes homelab, as schedulers automatically balance pods based on allocatable resources (node taints/labels also help). 8GB should handle 20-40 pods (e.g., Prometheus targets, Home Assistant), while 36GB can take heavier workloads like databases or future AI experiments without OOM issues. ...