D is the best match because AWS Lambda is a serverless compute service that runs code without provisioning or managing servers, which directly aligns with the requirement for minimal infrastructure and minimal operational support. Lambda supports Python runtimes, so the team can deploy the selected component as a microservice using the same language. For workloads with hundreds of requests per second, Lambda is designed to handle request-driven execution and automatically scales by running more concurrent instances of the function in response to incoming traffic, without the customer needing to pre-size or manage capacity.

The other options introduce significantly more operational responsibility. Option A still requires managing EC2 instances (AMI lifecycle, patching, capacity planning, scaling policies) and Spot interruptions can add complexity for a web microservice test. Option B (Elastic Beanstalk) reduces some management overhead but still relies on underlying instances and environment management (platform updates, scaling configuration, health monitoring), and it is not a serverless model. Option C (EKS with self-managed nodes) has the highest operational burden here: cluster and node management, scaling node groups, patching, and Kubernetes operations—this conflicts with “minimal operational support.”

Lambda fits the serverless microservices goal: deploy code as small independent services, integrate with other managed services (commonly API-driven invocation patterns), and rely on AWS-managed scaling and availability mechanisms rather than administering fleets or clusters.