I need to learn in depth EKS too, but I need to practice with new Gateway API on EKS and if I learn with k8s on my local computer, the real problems I can't learn well (production) and AWS is very expensive for me (not a company)

Honestly, not sure how much you can realistically learn on the free tiers.

@Null Shidare, I agree with Erik, the costs could be high for an individual, but you are both right, hands-on practice is the best way to learn. There is hidden complexity in any system, and especially in one as elaborate as AWS. I'd suggest practicing with ECS first, if you haven't yet. Fargate is fundamentally on-demand, and Fargate Spot is cheaper than the Savings Plans corporate customers buy. 😉 Obviously E*C*S has far fewer knobs and dials than EKS, but perhaps you can iron out some transferable IAM roles and policies; defining, storing and pulling a container for use on AWS (likely via EC*R*); and maybe some networking config. Also, do not hesitate to ask AWS for credits. If you have ties to a large AWS customer, the account manager might be able to help. If not, there are programs for new AWS users, including "cloud clubs", local user groups, and local one-day Summit conferences. Through these networks, you might meet an AWS representative to whom you could plead your case. Looking ahead, you could prepare this year to apply early next year for Community Builder status. Historically, Community Builders have received USD 500 in credits. If practicing on E*C*S first is helpful, I've demonstrated a pattern of flags in Terraform that lets you experiment in stages, turning on expensive stuff like PrivateLink endpoints, any containers or more containers, and in my example, Kafka, later, when you actually need it all. github.com/sqlxpert/docker-python-openapi-kafka-terraform-cloudformation-aws/tree/main/terraform You'll be able to implement something similar for EKS. Hope this helps, and please share what you build here, or join CloudPosse office hours on Wednesdays at 11:30 Pacific Time, via Zoom. It's a friendly community and people who know E*K*S much better than I ever will (I'm a simple-minded serverless or EC2 virtual machine person 🥲) can answer questions.

Hello, In my opinion, you can divide your learning into two parts: Kubernetes and AWS. You can learn Kube everywhere (on your laptop, AWS, GCP, OVH, etc.) so you can find a cheap Kube provider. And use AWS only for learning AWS stuff.