489 saved links, peaking in 2023 — the year of the great K8s learning wave in your saves. The recent corpus tells the next chapter: AI moving into the cluster, both as workload (LLMs on GKE/AKS) and as operator (K8sGPT, AI on-call). Per one of your saved reports, ML-on-Kubernetes job demand grew 180% while classic roles flattened.
Related: Containers & Docker · DevOps · Platform Engineering · Azure · GitOps
Start here: the right mental model
- “Let’s stop calling Kubernetes a container orchestrator” — the reframing that makes everything else click.
- How Kubernetes reinvented virtual machines — one of the most-shared K8s explainers ever, deservedly.
- Kubernetes is great — if you know what 90% of it not to use.
- “Kubernetes still feels weird” — what I wish I knew sooner (“wish I had it years ago” — your own note).
- architecture explained: etcd, kubelet, container runtime, component by component.
- the Borg papers — where it all comes from, 20 years on.
- Why does Kubernetes exist? (video).
AI meets the cluster
- AI as operator: K8sGPT — AI superpowers for Kubernetes SREs (+ the Backstage integration) · a ChatGPT bot that answers your Prometheus alerts · the kubectl OpenAI plugin.
- Intuit: GenAI for on-call across 325+ clusters and 7,000 services — the enterprise-scale case study closest to your incident-management platform.
- an intelligent alert-handling system with Prometheus, n8n, and OpenAI.
- AI as workload: deploying an agentic AI app on GKE with ADK and a self-hosted LLM · fine-tuning GenAI models on AKS with KAITO · private LLMs with NVIDIA NIM on a GPU node · one-checkbox Ray on GKE.
- scalable AI workflows: LLMs, Kubernetes, and multi-agent coordination · Kubernetes as the secret behind NVIDIA’s AI factories (video).
- the jobs report: platform engineering up, ML-on-K8s up 180%.
Running it in production
- the production checklist: best practices for SREs — the single most practical link in the tag.
- building a reliability foundation: crawl, walk, run.
- Observability: Prometheus + Grafana setup guide · a modern set of Grafana dashboards for K8s · K8s log monitoring with OpenTelemetry · SLOs for cluster resource utilization · Retina — cloud-agnostic network observability.
- Deploy patterns: deployment strategies: rolling, recreate, canary, blue/green · canary releases with Istio, Kiali, and the Gateway API · blue/green vs. canary, compared.
- Networking: how pod-to-pod communication actually works · understanding K8s networking end to end.
- Security: OWASP Kubernetes — prioritizing risks · Simulator: a K8s security-training platform that misconfigures clusters for you.
- the State of Kubernetes Cost Optimization report — tangible advice, per your note.
Kubernetes on Azure
- building a platform-engineering environment on AKS — the reference sample.
- Kubernetes in Azure: the 7 service options, compared.
- External Secrets + AKS integration · Azure DevOps self-hosted agents on K8s.
- AKS networking essentials and microservices CI/CD with Azure DevOps + AKS (videos).
- deploying intelligent apps with OpenAI on AKS (video).
Hands-on
- the official Minikube hands-on tutorial — still the canonical first cluster.
- 200+ Kubernetes labs and tutorials · free K8s labs · labs.play-with-k8s.com · helm-playground.com · k8sgames.com.
- iximiuz: learning Kubernetes by fixing failing pods and labctl — disposable K8s playgrounds from the CLI.
- CKA exam prep — plus the concepts the exam skips.
- the Kubernetes Resume Challenge — prove it in public.
- Home-lab route: Talos + K8s + Prometheus + Grafana on a Raspberry Pi 4 · why David runs K8s for personal stuff.
War stories & scale
- k8s.af — Kubernetes failure stories; the best ops education per hour spent.
- Kubernetes at Uber scale and Uber’s batch-compute resource story (videos).
- Agoda: moving 3 million CI jobs from VMs to Kubernetes.
- Monzo’s architecture lessons — what worked and what didn’t.
- Firefly: migrating from serverless to Kubernetes — the counter-current worth knowing.
- how the Kubernetes project coordinates 3,000 contributors.