We would like to request an enhancement to the Fivetran BigQuery connector to support connecting with least‑privilege, non‑admin service accounts, similar to what is supported by other data ingestion tools.

Currently, when attempting to configure a BigQuery connection, Fivetran requires elevated / admin‑level permissions. In our environment, granting admin‑level access is not permitted due to security and governance constraints.

For comparison, we successfully load data from BigQuery using Azure Data Factory with the same service account key and restricted permissions, without requiring admin access.

At present, our service account has:

• BigQuery Data Viewer (granted at the dataset level)

However, Fivetran additionally requires:

• BigQuery Job User (project level)
• BigQuery Read Session User (project level)

We are requesting that the BigQuery connector be enhanced to:

• Support connections using dataset‑level permissions only, where possible, or
• Clearly document and minimize required permissions in alignment with least‑privilege security principles, or
• Provide an alternative authentication/connection option that does not require project‑level roles when read‑only access is sufficient.

Reducing the permission requirements would enable teams operating in regulated or security‑constrained environments to adopt Fivetran without violating internal access controls.