AWS Certified Data Engineer Associate DEA-C01 Practice Question

Kinesis Data Firehose can synchronously invoke a Lambda function for each incoming batch before the data is delivered. The function can inspect each record, drop rows with missing product_id or invalid price, and then return the cleaned records to Firehose. The service automatically scales the Lambda concurrency and preserves the low-latency, serverless nature of the pipeline. Firehose cannot directly trigger an AWS Glue ETL job, so that option adds non-supported orchestration. Amazon S3 Object Lambda only transforms objects on GET requests; it does not alter data as it is written, so it cannot filter invalid records before storage. Using an S3 event-triggered Lambda cleans data after the object is already stored, doubles S3 PUT traffic, increases latency, and introduces additional buckets, making it less efficient than in-stream transformation.