It would be helpful to the backup team to get a consolidated offset/length so that all contiguous chunks of file are returned at once. Currently we return one granularity-sized offset/length pair at a time. So if we have a file with a 1Mb granularity and many of the bytes in the first 3Mb have changed, WT currently returns 3 cursor items. The first for offset 0/1M length, the seconds for offset 1M/1M length and then the third for offset 2M/1M length.
Cloud would prefer an API for this functionality, default off, to retain backward compatibility with current code. Otherwise this probably would be an always-on feature instead of an API.
The API will look like incremental=(consolidate=true|false) on the configuration of the primary backup cursor for incremental.