The client must manage the watermark and send it as input while initiating the bulk claim data API sync.
BulkData ingestion in Data Cloud has limitations in processing data into Data Cloud parallelly. New sync requests are accepted from the client, queued, and then dispatched sequentially after the previous job is fully complete.
Only one job can be created for a particular object. An additional job for the same object type can’t be created until the job is fully processed.
The number of concurrent requests or jobs allowed is 5.