Registers a new customer import or update batch and returns a pre-signed S3 URL that the data file should be uploaded to.

The bulk workflow is:

1. Call this endpoint with the batch name, file name, and mode. The response carries the batchId and signedUrl.
2. PUT the CSV or XLSX file to signedUrl. Uploading the file is what triggers processing, so the batch remains pending until the file arrives.
3. Poll Get import batch until status is completed, completed_with_errors, or failed.
4. If the batch completed with errors, fetch the annotated error file with Get import error file, then correct and resubmit those rows using Retry import batch.

File constraints: the file must be a .csv or .xlsx file, may not exceed 10 MB, and the file name may not contain a forward slash ( / ), backslash ( ** ), or double period ( .. ).