AWS Lake Formation provides centralized data access control, including fine-grained (column-level) permissions for data stored in S3 and accessed through services like Amazon Athena.
Using a Lake Formation blueprint to ingest data from Aurora MySQL into the data lake keeps ingestion and governance integrated. When QuickSight uses Athena as the data source, Athena enforces Lake Formation's column-level permissions automatically. This allows the marketing team to see only the authorized subset of columns without custom access-control logic.
Options A, B, and C rely on manually limiting columns at ingestion time or using IAM or S3 bucket policies, which do not provide true column-level authorization for SQL queries and require significantly more manual work and maintenance.

Amazon SQS is the best messaging layer here because it providesdurable bufferingand decouples microservices so they can scale independently while absorbing bursts without losing data. AWS recommends SQS for decoupled architectures where producers and consumers operate at different speeds. For compute, AWS Fargatereduces operational overhead because AWS manages the underlying servers, patching, and capacity layer. API Gateway and WebSockets are not the primary answer for internal asynchronous microservice communication, and SNS is fanout-oriented rather than the most natural queueing model for durable work processing between services. Therefore, SQS plus ECS on Fargate is the most appropriate low- overhead design.

Using API Gateway and Lambda enables serverless handling of form submissions with minimal cost and infrastructure. When coupled with Amazon SNS, it allows instant email notifications without running servers, making it ideal for low-traffic workloads.
Reference: AWS Documentation - Serverless Contact Form with API Gateway, Lambda, and SNS

This solution offers the least operational overhead while meeting the security and global availability requirements:
Amazon CloudFront and S3: Hosting the static frontend on S3 and serving it via CloudFront provides low- latency global distribution, high availability, and security. S3 is a cost-effective and serverless option for hosting static assets, and CloudFront ensures that the application is cached closer to the users, reducing latency globally.
AWS Lambda and API Gateway: Refactoring the microservices to use Lambda functions with API Gateway allows for a fully serverless, scalable, and highly available backend. This reduces the need for managing EC2 instances, as Lambda automatically scales to meet demand and only charges for the actual usage.
* RDS for MySQL: Migrating the MySQL database from an EC2 instance to Amazon RDS significantly reduces operational overhead. RDS manages backups, patching, and scaling, and it offers high availability options (e.g., Multi-AZ).
Option A and D involve using EC2 Reserved Instances for the database, which requires more operational maintenance than using RDS.
Option C suggests using a Network Load Balancer with Lambda, which adds unnecessary complexity for this use case.
AWS References:
Amazon S3 and CloudFront Integration
AWS Lambda with API Gateway
* Amazon RDS for MySQL

The correct answer isAbecause the application is alightweight data processing toolthat always completes within3 minutes, which fits well within the execution model ofAWS Lambda. Lambda is a fully managed compute service that eliminates the need to provision or manage servers, clusters, operating systems, or container hosts. Since the company is willing to make any required changes to the image, packaging the workload as aLambda container imageis the option with the least operational overhead.
Lambda is well suited for short-lived, event-driven, or on-demand processing jobs. It automatically scales, charges only for execution time, and removes nearly all infrastructure administration. Compared with container orchestration services, Lambda requires less setup, less monitoring of infrastructure, and no task or cluster capacity planning.
Option B is incorrect becauseAmazon EKS with Fargatestill requires Kubernetes management constructs and introduces more operational complexity than Lambda. Option C is simpler than EKS, butAmazon ECS with Fargatestill involves managing task definitions, services, and networking, which is more overhead than Lambda for a brief processing job. Option D is incorrect because using a custom AMI with AWS Batch introduces more compute management and is unnecessary for a very short, lightweight workload.
AWS guidance favorsserverless serviceswhen the workload is short-lived and infrastructure management should be minimized. Because this job always completes quickly and the company wants the least operational overhead,AWS Lambda with a container imageis the best solution.