At the moment, Dataverse 'Incremental Update' Dataverse snapshots are created to a strict time schedule, configured when then synapse link is setup.

Also, snapshots are also only created (and able to be consumed) upon the arrival of new data, so old data can be handing around in a snapshot until new data arrives.

There are legitimate use-cases to manually stipulate a point of time for a snapshot creation, namely:

• Gathering of data at an arbitrary timepoint determined by an extraneous process, e.g. batch / scheduled processes.
• Forcing a cut of data at a time where new data is not arriving.
• Reducing cloud costs associated with ADF pipeline dataflows (explicit snapshot creation reduces the required frequency of the dataverse update, and inherent use of Integration Runtime cores).