My aim is to back up and restore a dynamodb table to/from S3. I'm aware about the Import/Export using data pipelines and DynamoDb streams. My concern using Data Pipeline is that when the backup process is complete, I cannot be sure as to what 'point-in-time' the backup was performed. For e.g. if the backup process started at 11am and ended at 1130am, the items in my backup could be for anything that happened in the dynamodb table from 11am to 1130am.
How can I use DynamoDb streams to compliment the data pipeline ?
How can I restore the dynamodb table to a particular 'point-in-time' ?