Which solution meets these requirements with the LEAST operational overhead?
Configure point-in-time recovery for the table.
Use AWS Backup for the table.
Use an AWS Lambda function to make an on-demand backup of the table every hour.
Turn on streams on the table to capture a log of all changes to the table in the last 24 hours. Store a copy of the stream in an Amazon S3 bucket.
Explanations:
Point-in-time recovery (PITR) in DynamoDB allows for restoring a table to any point within the last 35 days (24 hours in this case). This solution requires minimal operational overhead and provides a simple, built-in mechanism for recovery without the need for additional management.
AWS Backup provides centralized backup management, but it requires more configuration and monitoring compared to PITR. It may involve higher operational overhead and isn’t the most efficient solution for the specific 24-hour window recovery requirement.
Using a Lambda function to make on-demand backups every hour introduces significant operational overhead. It requires custom code, scheduling, and management of backup frequency, which is more complex than leveraging PITR.
Enabling streams captures changes to the table, but it does not allow for point-in-time table restoration. Storing streams in S3 requires additional management and cannot easily revert the table to a specific state within the last 24 hours. It adds complexity without meeting the exact requirement for easy restoration.
In my experience, the answer is:
Configure point-in-time recovery for the table.
My best guess is:
Configure point-in-time recovery for the table.