Christian Huth's Tech Blog

Sign in Subscribe

Reducing Pod Startup Times - Part 1: Running an Image Pool

Reducing Pod Startup Times - Part 1: Running an Image Pool

A pod can sit in ContainerCreating for minutes on a node that has been up for weeks. It has capacity. It just doesn't have the image. Part 1: keeping an image pool warm on every node, and where that breaks.

Too Long, Didn't Apply: Server-Side Apply for Oversized CRDs in Argo CD ApplicationSets

Too Long, Didn't Apply: Server-Side Apply for Oversized CRDs in Argo CD ApplicationSets

TemplatePatch is one of Argo CD's lesser-known features. It is also the cleanest way to enable server-side apply for exactly one generated Application - and to stop oversized CRDs from failing to sync.

Migrating a Bitnami PostgreSQL Helm Release to CloudNativePG

Migrating a Bitnami PostgreSQL Helm Release to CloudNativePG

The Bitnami catalog moved most images to bitnamilegacy and stopped updating them. CloudNativePG can bootstrap a new cluster straight from the running Bitnami instance — here's the walkthrough.

The Argo CD Ownership Conflict — A kube-green Case Study

The Argo CD Ownership Conflict — A kube-green Case Study

Any controller that writes to a field Argo CD also manages — sleep schedulers like kube-green, autoscalers, cert rotation, whatever — collides with self-heal the same way. Five ways to solve it, via a kube-green case study.

A Highly Available Egress Gateway for Kubernetes with Cilium and kube-vip

A Highly Available Egress Gateway for Kubernetes with Cilium and kube-vip

Cilium's Egress Gateway gives your pods a static, routable identity for outbound traffic — but it won't manage that IP for you. Here's how to combine it with kube-vip to get a dynamic, BGP-announced, highly available egress IP.

Consolidating to Cilium: Replacing kube-proxy, MetalLB, and ingress-nginx in Three Phases

Consolidating to Cilium: Replacing kube-proxy, MetalLB, and ingress-nginx in Three Phases

In March 2026, ingress-nginx reached end-of-life. For clusters already running Cilium as their CNI, this is the right moment to go all-in: replace kube-proxy, MetalLB, and ingress-nginx with Cilium's built-in equivalents — ending up with a single networking component doing the work of four.

Mastering Vim: A Cheatsheet Guide for Fast & Efficient Text Editing

Mastering Vim: A Cheatsheet Guide for Fast & Efficient Text Editing

Vim is a powerful, keyboard-driven text editor that excels in speed and efficiency. This cheatsheet covers essential commands for navigation, editing, and file management—perfect for beginners or anyone looking to boost their productivity in Vim.

Reviving a Kubernetes Cluster After Long-Term Downtime: Expired Certificates Edition

Reviving a Kubernetes Cluster After Long-Term Downtime: Expired Certificates Edition

Learn how to recover a Kubernetes cluster after long-term downtime due to expired certificates. This step-by-step guide covers control plane restoration, worker node rejoining, and tips to prevent future issues.

Docker Hub Rate Limits Are Coming – Here’s How to Outsmart Them with a private Registry

Docker Hub Rate Limits Are Coming – Here’s How to Outsmart Them with a private Registry

Avoid Docker Hub rate limits by using Harbor as a proxy cache. Learn how to install and configure Harbor, redirect image pulls in Kubernetes, Docker, Podman, and OpenShift, and reduce the costs of Docker Hub subscriptions with IDMS/ITMS, admission webhooks and CoreDNS rewrites.

Enhance Argo CD Observability: A Step-by-Step Guide to Integrating Prometheus Metrics into the UI

Enhance Argo CD Observability: A Step-by-Step Guide to Integrating Prometheus Metrics into the UI

Bringing more observability to the Argo CD UI

Verifying Helm Charts for improved security

Verifying Helm Charts for improved security

Helm Charts are the go-to way of packaging and sharing applications in