Understand how to work with S3 Versioning in the context of restoring files.
How will this affect storage?
Enable S3 Versioning on UAT, verify with tests that the versioning works.
Verify that a previous version can be retrieved (even if stored in deep glacier).
We may need a standard way to retrieve these files (script or code that can retrieve a targeted granule as needed)
If everything on UAT checks out, do this in PROD as well and test it there.