Learning Kubernetes from Scratch
LinkedIn post by yetmike — view post · ~1,532 reactions · 63 comments · 202 reposts
The premise
“If I had to learn Kubernetes from scratch today… I’d ignore 90% of the tutorials online.” No weeks reading docs before touching a cluster; no YouTube playlist that ends at kubectl get pods. Learn from the right sources instead — the fastest path into production-grade Kubernetes isn’t on Udemy.
The 13 bookmarks
(Numbered as posted; his list visibly jumps 12 → 15 — nothing labeled 13/14 appears in the post. In the original they’re written with _ instead of . to dodge link scrapers, with instruction “replace _ with .”)
| # | Resource | Why he picks it |
|---|---|---|
| 1 | killer.sh | Hardest CKA/CKAD/CKS simulator available; closest thing to the real exam |
| 2 | k8s.af | Real Kubernetes failure stories from real teams — more useful than any course |
| 3 | roadmap.sh/kubernetes | The full skill map, so you know what you don’t know yet |
| 4 | kubernetes-the-hard-way | Bootstrap a cluster by hand once; you’ll never forget what the control plane does |
| 5 | cncf.io/projects | Every graduated and incubating project ranked by maturity — know the ecosystem |
| 6 | prometheus.io/docs | Observability is not optional — start here before you think you need it |
| 7 | grafana.com/tutorials | Dashboards that surface problems instead of looking pretty |
| 8 | opentelemetry.io/docs | The standard — stop learning vendor-specific instrumentation |
| 9 | fluxcd.io/docs | GitOps done right; understand this before you touch Argo |
| 10 | argoproj.github.io | Progressive delivery, rollouts, workflows — one org, four tools |
| 11 | trivy.dev | Scan images, configs, and IaC in one tool — “no excuse not to” |
| 12 | falco.org | Runtime threat detection — the thing you wish you had after an incident |
| 15 | helm.sh/docs | You’ll use it whether you like it or not — learn it properly |
Closing line: “Sometimes changing where you learn matters more than logging another hour in a course you already plateaued on.”
Notable replies
- Oleksandr Sergienko (Infra Architect): the list covers tooling but misses Day-2 operational maturity — the bridge from tutorials to production. Unglamorous realities: debugging an etcd latency spike under disk I/O pressure, sizing control-plane components before the API server drops requests during deploy surges, and knowing when NOT to use Kubernetes (managed serverless or plain VMs can win 10× ROI). His question back: which production incident taught you more than all these docs combined?
- Krishasto B.: agrees partially — add performance/load/endurance testing so you find the component that bottlenecks (fine at 20 concurrent requests, slow at 60).
- Mohit Mishra (Cloud Architect): “Build your own playground instead” — e.g. buy an affordable domain and self-host (comment truncated in capture).
Same author’s related post: Yetmike — Golden Kubestronaut (order-of-exams strategy across all 16 certs).