How to transfer expired DynamoDB items to S3 with least operational overhead?

Understanding DynamoDB TTL and Streams Integration

When DynamoDB Time to Live (TTL) is enabled, items that pass their expiration timestamp are automatically deleted by a background process. Crucially, these deletions generate REMOVE records in DynamoDB Streams, which allows downstream consumers to react to expired items in real time.

Why Option B is Correct

Option B — configuring DynamoDB Streams to invoke an AWS Lambda function — is the correct answer because:

• DynamoDB Streams natively captures TTL-driven deletions as REMOVE events.
• AWS Lambda is a fully managed compute service, meaning zero server provisioning, zero patching, and automatic scaling.
• The 5-minute processing window is well within Lambda's maximum execution time (15 minutes).
• This approach delivers the least operational overhead among all options.

Why the Other Options Are Wrong

• Option A (DAX): DynamoDB Accelerator (DAX) is an in-memory cache designed to reduce read latency. It cannot be "configured to query for expired items based on TTL" in a meaningful way for archival, and it certainly does not push data to S3. This is a distractor.

• Option C (ECS on EC2): Running a custom application on Amazon ECS with EC2 instances requires managing underlying EC2 infrastructure, container orchestration, scaling policies, and patching. This introduces significant operational overhead, directly contradicting the requirement.

• Option D (EventBridge rule): Amazon EventBridge can schedule Lambda invocations, but it does not receive DynamoDB TTL expiration events directly. You would need a custom polling mechanism or a scheduled scan, which is far less efficient and more operationally complex than using DynamoDB Streams.

Key Takeaway

Always pair DynamoDB TTL with DynamoDB Streams + Lambda when you need to react to item expirations. This pattern is serverless, event-driven, and minimizes operational overhead.