The correct answer is: Use an AWS Lambda function.

AWS Lambda is a serverless, event-driven compute service ideal for workloads with irregular traffic patterns. It automatically scales with demand, charges only for the time your code runs, and integrates natively with API Gateway. Since the background tasks complete within a few seconds and requests are sporadic, Lambda provides a highly cost-effective and operationally efficient solution.

Why the other options are incorrect:

Using an AWS Glue job is inappropriate for this use case. AWS Glue is primarily used for ETL workloads and data transformation tasks that operate on structured data sources, not for near-real-time API-backed compute processing. It also has longer startup times and is not designed for latency-sensitive interactions.

Deploying a containerized service on Amazon EKS is overkill for this scenario. EKS requires managing a Kubernetes control plane and worker nodes, leading to unnecessary complexity and costs, especially for an infrequent workload.

Hosting containers on Amazon ECS with EC2 also introduces fixed infrastructure costs since EC2 instances run continuously unless manually scaled down. This would result in idle resource charges during periods of no traffic, making it less cost-effective compared to a fully serverless model.

Why this question matters: It evaluates your understanding of how to design cost-efficient, event-driven architectures using AWS serverless services like Lambda and API Gateway. It also helps assess your ability to match unpredictable workloads with scalable compute options.

Tip: For unpredictable workloads where compute is needed only on demand and with short execution time, AWS Lambda is typically the optimal solutionespecially when integrated with API Gateway.

Full AWS Practitioner Certification Question