Improve Response Archiver to Handle Large Volumes Efficiently

### Description:

Recent load tests have highlighted a limitation with our current response archiver implementation. When the database contains a large number of responses, the existing Lambda function may fail to process all items within its 15-minute execution window. This can result in responses that are ready to be archived remaining in the database indefinitely.

### Proposed Solution:

We should explore a new approach leveraging DynamoDB features to make archiving more reliable and scalable:

- DynamoDB TTL: Configure TTL on responses (including actual data and confirmation codes) to automatically expire items after a defined number of days.
- DynamoDB Streams: Use DynamoDB Streams to detect when a response is deleted due to TTL expiration.
- Archival Lambda: Connect a Lambda function to the stream to archive deleted responses to S3, including attachments if necessary.

Optional Enhancement:

To reduce risk, we could introduce an intermediate Lambda that pushes deleted responses to an SQS queue before archival, providing an extra layer of reliability and decoupling.

### Motivation / Benefits:

- Ensures responses are archived reliably even under heavy load.
- Removes dependency on a single long-running Lambda process.
- Leverages native DynamoDB features for efficiency and scalability.

Start by locating the current response archiver and reviewing the load-test behavior described in the issue. Map how responses, confirmation codes, and attachments are stored and archived, then evaluate a DynamoDB TTL/Streams design, with SQS as an optional reliability layer; done means reliable archival under large volumes without the current execution-window failure.

Written by the indexing model from the issue text.