Tableau connection

Connect a task to Tableau Server or Tableau Cloud to refresh extracts, publish workbooks, or query metadata over a managed Leoflow Connection. TableauHook signs in to the REST API.

Declare the provider

# leoflow.yaml
dag_id: tableau_signin
connectors:
  - tableau

URI shape

tableau://<login>:<password>@<host>/<site>

The control plane builds this from the Connection’s fields and exports it as AIRFLOW_CONN_<CONN_ID>. The site id is carried in Schema; the password is percent-escaped.

Fields the UI asks for

Example DAG

# dag.py
from airflow.sdk import DAG, task


@task
def list_workbooks():
    from airflow.providers.tableau.hooks.tableau import TableauHook

    with TableauHook(tableau_conn_id="tableau_default") as hook:
        workbooks = hook.get_all(resource_name="workbooks")
        print("workbook count:", len(list(workbooks)))


with DAG("tableau_signin", schedule=None, catchup=False, tags=["example"]):
    list_workbooks()

# leoflow.yaml
schema_version: "1.0"
dag_id: tableau_signin
python_version: "3.12"
connectors:
  - tableau
connections:
  - tableau_default

Security notes

• Prefer personal access tokens over username/password: put token_name + personal_access_token in Extra. PATs can be revoked independently and don’t expose the account password.
• TLS: Tableau Server should be HTTPS; the host must present a valid cert.
• Never log AIRFLOW_CONN_TABLEAU_DEFAULT; it carries the password.

Related

• ADR 0019 — secret encryption at rest.
• ADR 0021 — agent secret delivery (AIRFLOW_CONN_<CONN_ID>).
• TestTableauConnectionURIShapeIntegration — chain-of-custody delivery test.