Kube Builders

Mac Chaffee challenges the assumption that scaling means going distributed by advocating for "one big box" infrastructure. He explains that while vertical scaling on powerful single machines can take you far and simplifies operations significantly, the critical challenge remains: how do you solve failover and high availability? He observes a growing trend where companies deploy single-node Kubernetes for on-premise products, recognizing that Kubernetes solves numerous application runtime problems beyond just distributed computing. Watch the full episode: https://ku.bz/9nFPmG85f

This tutorial shows how to advertise Kubernetes LoadBalancer IPs to a LAN by configuring Cilium BGP Control Plane, an IP pool, Cilium BGP resources, and FRR on a UniFi Gateway. More: https://ku.bz/y36nr3n9d

Federico Iezzi, Customer Engineer at Google Cloud, explains how his team achieved 1 million output tokens per second using Qwen 3.5 27B, vLLM, GKE Autopilot, and NVIDIA B200 GPUs. You will learn: - Why memory bandwidth limits decode performance - How Federico chose between tensor and data parallelism - What changed after enabling multi-token prediction and reducing the KV cache footprint with FP8 quantization Watch (or listen to) it here: https://ku.bz/1xD9Md0mb 🌟 This episode is brought to you by LearnKube. Download the free book, The Technical Guide to Kubernetes Rightsizing, to understand what Prometheus and Grafana cannot tell you about safely reducing requests and limits: https://learnkube.com/kubernetes-rightsizing With @Birthmarkb

Grafana Beyla is an open source eBPF-based tool that auto-instruments HTTP/gRPC apps to collect essential traces and RED metrics and exports them in OpenTelemetry or Prometheus formats. More: https://ku.bz/61N3-25F4

Rūsternetes is a from-scratch Rust reimplementation of Kubernetes with its own: - API server, - scheduler, - controller manager, - kubelet, - kube-proxy, - controllers, - and conformance tests. More: https://ku.bz/yjgm-qPVh

Marc Campora shares his perspective on the hybrid approach of running AWS Lambda on Kubernetes. He explains why this strategy, while potentially helpful for bin packing optimization and utilizing idle capacity, ultimately misses the point because you still inherit all of Kubernetes' operational complexity - deployment, resizing, and operations. Instead, Marc advocates for a mixed architecture approach where you run Kubernetes for services needing containers alongside native Lambda functions in AWS or Azure environments. The key is implementing proper abstraction layers through service bus or service mesh technologies, making the underlying execution platform transparent to your applications. Watch the full episode: https://ku.bz/5gMTkzLhV

This case study explains how one private container registry was shared across six AKS clusters in three regions. It covers ACR geo-replication, image pull reliability, network access, permissions, and operational breakages. More: https://ku.bz/CjrRY3Jp8

NVIDIA AICR provides optimized, validated, and reproducible GPU runtime tooling for Kubernetes, with recipes for deploying training and inference stacks on NVIDIA-accelerated clusters. More: https://ku.bz/D355D3SW7

This article explains how to control runaway GPU spend in Kubernetes by adding taints, quotas, labels, Prometheus rules, and admission controls so teams can see who is using expensive GPU workloads and why. More: https://ku.bz/6cFxnhHGb

This week on Learn Kubernetes Weekly 193: 🔍 Which of our Containers are Chainguard? 💸 One Forgotten Notebook on an A100. $1,800 a Month. 🔄 How an Admin Cluster Keeps Application Clusters in Sync with GitOps 📉 Cost Optimization of Spark on Kubernetes Batch Workloads on Public Clouds 🚪 ingress-nginx Is Archived: How We Migrated to kgateway (and Didn't Break Prod) Read it now: https://kube.today/issues/193 ⭐️ This newsletter is brought to you by LearnKube — master Kubernetes with hands-on training designed for engineers who want to learn the smart way https://ku.bz/hypSbyc-V

VictoriaMetrics MCP Server connects AI tools to VictoriaMetrics APIs and documentation for metric queries, alert analysis, observability debugging, automation, and read-only monitoring workflows. More: https://ku.bz/Q_chvmrhf

This tutorial explains why standard GKE Ingress breaks under Istio STRICT mTLS and shows how to replace it with an Istio Ingress Gateway, Gateway resource, and VirtualService. More: https://ku.bz/lNmNzN4HW

This tutorial explains how to build a Kubernetes operator that automatically manages Synology DSM reverse proxy rules from annotated Services, Ingresses, and Argo CD Applications. More: https://ku.bz/P015Vw0wn

Zain Malik, Software @ Exostellar, explains how his team doubled the number of managed Kubernetes clusters without growing headcount. He describes how automation through Cluster API and GitOps transformed their operations by reducing firefighting and manual node pool upgrades. The improvements allowed them to reduce from multiple dedicated engineers to just one person handling cluster upgrades, with the ability to rotate different team members through standardized processes. Zain highlights how removing human intervention from critical operations and implementing defined rules for workload disruption significantly reduced user-impacting incidents during maintenance windows. Watch the full episode: https://ku.bz/5PLksqVlk

Migratowl is an AI-powered dependency migration analyzer that upgrades dependencies in isolated sandboxes, runs tests, reads changelogs, explains breakages, and suggests fixes. More: https://ku.bz/5Ddt6Pjj1

Vitalii Horbachov, Staff Software Engineer at Agoda, explains the massive scale challenges his team faced when running iOS CI/CD infrastructure. With 20,000 iOS tests requiring 200 Mac Minis and pipelines taking hours to complete, Agoda turned to Kubernetes as a job scheduler to manage workloads at scale. However, this approach introduced significant problems that ultimately led them to develop a custom solution for integrating native macOS workloads with Kubernetes - setting the stage for their innovative macOS VZ Kubelet project. Watch the full episode: https://ku.bz/q_JS76SvM

dt is a set of utilities available as a standalone or Helm Plugin for making offline work with Helm charts easier. You can create reproducible and relocatable packages for Helm charts that can be easily moved across registries without hassles. More: https://ku.bz/bSVHSvXqj

Cardamon audits Prometheus metrics against query logs, rules, and Grafana dashboards, then finds unused series and generates drop relabeling rules to reduce TSDB storage and cardinality waste. More: https://ku.bz/Hg6vtcg09

"He'd been looking forward to in-place pod resizing for years." Pedro Ignácio discusses Karpenter's limitations and the need for multi-cloud support. While Karpenter works well for AWS, Pedro emphasizes that global enterprises need cloud-agnostic tools that work across Azure, Google, and other platforms to standardize workflows and avoid vendor lock-in. Watch the full interview: https://ku.bz/515x11H-T This interview is a reaction to Kensei Nakada's episode https://ku.bz/bRd0243xQ

Three tools Rodrigo Bersa watches closely at AWS: 1. Karpenter — node autoscaler that consolidates capacity fast and spins up nodes without waiting. 2. KRO (Kubernetes Resource Orchestrator) — define resource graph definitions that provision Kubernetes and infrastructure resources together. Pair it with ACK for AWS-managed infra in the same lifecycle. 3. Cedar — an upcoming RBAC extension for Kubernetes 1.36/1.37 that adds deny and conditional permissions. Today, RBAC is additive-only. Cedar changes that. Each solves a different layer of the platform problem. Watch the full interview: https://ku.bz/dB7PDNt0v