A company stores sales data in a DynamoDB table with DynamoDB Streams enabled. Each item has a TransactionStatus attribute with values failed, pending, or completed. The company wants to be notified for failed sales where Price exceeds a threshold. What approach requires the least development effort to set up this notification?

The correct option leverages Lambda event filtering, which allows you to specify criteria for invoking the Lambda function directly from the DynamoDB Stream. This significantly reduces the amount of code needed within the Lambda function itself, as the filtering logic is handled by the AWS service. The Lambda function then only needs to publish to SNS when invoked, making it the least development effort. The second option requires writing the filtering logic inside the Lambda function, which is more development effort than using built-in filtering. The third option is incorrect because DynamoDB Streams cannot directly publish to an SNS topic with complex event filtering; an intermediary like Lambda is required. The fourth option is incorrect because CloudWatch Alarms monitor metrics, not the content of DynamoDB Stream records, and cannot perform the necessary content-based filtering.