A company collects customer satisfaction surveys on its website. Surveys can reach thousands of customers per hour. Results are currently emailed for manual review. The company wants to automate the workflow and keep survey results available for the previous 12 months. Which solution meets these requirements in the MOST scalable way?

The correct solution leverages a fully managed, serverless architecture that scales automatically. Amazon API Gateway provides a scalable entry point for the survey data. Amazon SQS acts as a buffer, decoupling the ingestion from processing and handling potential spikes in survey submissions. An AWS Lambda function processes messages from SQS, calls Amazon Comprehend for sentiment analysis (a service specifically designed for text analysis), and stores the results in Amazon DynamoDB, a highly scalable NoSQL database. DynamoDB's Time-to-Live (TTL) feature automatically purges records after 365 days, meeting the retention requirement efficiently. The option using an EC2 instance introduces operational overhead and less scalability compared to serverless. The S3-based option is less suitable for real-time processing of individual survey results and incorrectly suggests Amazon Rekognition for sentiment analysis (Rekognition is for image/video analysis, not text). The option using Amazon Lex for sentiment analysis is incorrect; Lex is for building conversational interfaces, not general sentiment analysis of text.