A business-critical application runs on Amazon EC2 instances and stores data in an Amazon DynamoDB table. The company must be able to revert the table to any point within the last 24 hours. Which solution meets this requirement with the LEAST operational overhead?

Configuring point-in-time recovery (PITR) for a DynamoDB table is the correct solution because it allows recovery to any point within the last 35 days, including the required 24 hours, with minimal operational overhead. It's a single setting that enables continuous backups. AWS Backup can back up DynamoDB tables, but PITR is specifically designed for continuous, granular recovery within a defined window, making it more suitable for this requirement. Using a Lambda function for hourly on-demand backups would incur significant operational overhead to develop, maintain, and monitor the function, and it would only allow recovery to hourly intervals, not "any point." DynamoDB Streams capture changes but require additional logic and storage (like S3) to reconstruct the table state, adding complexity and operational overhead compared to PITR.