Question 20 – Category: CSAP – Design for New Solutions

[seth_e]

In this question, either Lambda or Fargate could be correct. The explanation chooses Lambda because “Running a Fargate cluster continuously is more expensive than running Lambda functions which only run on-demand.” However, the question doesn’t specify the traffic pattern. The word “unpredictable” describes variance, not volume—traffic can be unpredictable while still being continuous. If the workload has sustained traffic without idle periods, Fargate may actually be more cost-effective since Lambda’s per-invocation pricing becomes expensive at scale.

===

20 Question Category: CSAP – Design for New Solutions

A company wants to create a new service that will complement the launch of its new product. The site must be highly-available and scalable to handle the unpredictable workload, and should also be stateless and REST compliant. The solution needs to have multiple persistent storage layers for service object metadata and durable storage for static content. All requests to the service should be authenticated and securely processed. The company also wants to keep the costs at a minimum. Which of the following is the recommended solution that will meet the company requirements?

With the above solutions, the correct answer is: Configure Amazon API Gateway with the required resources and methods. Create unique Lambda functions to process each resource and configure the API Gateway methods with proxy integration to the respective Lambda functions. Control user access to the API by using Amazon Cognito user pools. Store service object metadata in an Amazon DynamoDB table with Auto Scaling enabled. Create a secured Amazon S3 bucket to store the static content. Generate presigned URLs when referencing objects stored on the S3 bucket.

The option that says: Package the REST service on a Docker-based container and run it using the AWS Fargate service. Create a cross-zone Application Load Balancer in front of the Fargate service. Control user access to the API by using Amazon Cognito user pools. Store service object metadata in an Amazon DynamoDB table with Auto Scaling enabled. Create an encrypted Amazon S3 bucket to store the static content. Generate presigned URLs when referencing objects stored on the S3 bucket is incorrect. Running a Fargate cluster continuously is more expensive than running Lambda functions which only runs on-demand.

[JR-TutorialsDojo]

Hello seth_e,

Thanks for sharing your thoughts on this item.

You’re absolutely right that in real-world scenarios, both Lambda and Fargate could be valid depending on traffic patterns. If the workload is continuous and high-volume, Fargate may indeed be more cost-effective, while Lambda shines when traffic is bursty or unpredictable with idle periods.

To clarify, unpredictable workloads refer to traffic patterns that fluctuate without a consistent baseline, such as sudden spikes, seasonal surges, or irregular request volumes that are difficult to forecast. In these cases, provisioning fixed compute capacity can lead to either over-provisioning (wasted cost) or under-provisioning (availability risks). Lambda’s event-driven scaling model is designed to handle exactly this kind of variability, automatically adjusting capacity without upfront planning.

Additionally, the scenario explicitly emphasizes requirements such as being stateless, REST-compliant, highly available, scalable for unpredictable workloads, and cost-efficient. These keywords strongly align with Amazon API Gateway and AWS Lambda, which is why that option is considered the best fit in this case.

I hope this helps! Let us know if you need further assistance.

Regards,
JR @ Tutorials Dojo