Design High-Performing Architectures Question

[Diogo Simoes]

I have some concerns/doubts about this one:

A new online banking platform has been re-designed to have a microservices architecture in which complex applications are decomposed into smaller, independent services. The new platform uses Kubernetes, and the application containers are optimally configured for running small, decoupled services.

The new solution should remove the need to provision and manage servers, let you specify and pay for resources per application as well as improve security through application isolation by design.

Which of the following is the MOST suitable solution to implement to launch this new platform to AWS?

The answer considered correct is “Use AWS Fargate on Amazon EKS with Service Auto Scaling to run the containerized banking platform”.

Another available option is “Use Amazon ECS to run the Kubernetes cluster on AWS Fargate”.

Problems I see with this answer and the explanation:

– “Service Auto Scaling” does not exist in K8S. It exists, however, in ECS.

– The explanation goes into “AWS Fargate is a serverless compute engine for containers that works with both Amazon Elastic Container Service (ECS) and Amazon Elastic Kubernetes Service (EKS).” but never explains why EKS is picked instead of ECS

Why would EKS be picked instead of ECS in this case?

IMO, EKS makes sense because the bank would, in theory, already have k8s expertise. But then the EKS answer throws “Service Auto Scaling” in the mix, which is quite out of place and seems wrong, as I’ve explained above.

Cheers

• This discussion was modified 6 months, 3 weeks ago by  Diogo Simoes.

[obmon]

The reason EKS on Fargate is chosen over ECS on Fargate is mainly because the scenario explicitly mentions that the new platform is based on Kubernetes. If the organization already has Kubernetes workloads, tooling, and expertise, EKS allows them to reuse their existing ecosystem and avoid retraining or rewriting everything for ECS.

Your observation is correct: “Service Auto Scaling” is not a Kubernetes concept — that part of the answer is inaccurate. In Kubernetes you would use HPA/VPA, not ECS-style Service Auto Scaling. The exam answer is conceptually pointing to “autoscaling of services on EKS,” but the terminology is wrong.

EKS is picked because the bank uses Kubernetes; Fargate removes node management. The scaling wording is just an imprecise part of the test.

• This reply was modified 5 months ago by  obmon.

[JR-TutorialsDojo]

Hello Diogo Simoes,

Thank you for taking the time to share your feedback and concerns. We really appreciate the detailed observations.

Using Amazon ECS to run a Kubernetes cluster on AWS Fargate is not appropriate, as ECS is primarily designed for orchestrating Docker containers rather than Kubernetes.

Regarding Service Auto Scaling, you are right that this terminology applies to ECS. In Kubernetes/EKS, scaling is typically achieved through mechanisms such as the Horizontal Pod Autoscaler (HPA) or the Cluster Autoscaler. We recognize that the wording in the explanation could cause confusion, and we will make the necessary updates to ensure the distinction is clear and technically accurate. These changes should be reflected on the portal soon.

Thanks again for pointing this out. Let us know if you need further assistance.

Best regards,
JR @ Tutorials Dojo