Apache Airflow Tableau Provider

5.3.4 · active · verified Sun Apr 12

The Apache Airflow Tableau Provider enables seamless integration between Apache Airflow and Tableau Server/Online. It offers operators and hooks to automate various Tableau-related tasks, such as refreshing data sources, publishing workbooks, and managing server content, directly within Airflow DAGs. The provider is actively maintained as part of the Apache Airflow project, with regular releases aligning with Airflow's provider release cycles, currently at version 5.3.4.

Warnings

deprecated Authentication by personal access token (PAT) is officially deprecated in provider version 2.1.0 and later due to concurrency issues. Tableau automatically invalidates PATs if multiple parallel connections use the same token, leading to job failures.
Fix: Migrate to JSON Web Token (JWT) authentication for robust and scalable Airflow-Tableau integrations, especially in environments with parallel tasks.
breaking Provider version 4.0.0 removed deprecated classes paths, specifically `tableau_job_status` and `tableau_refresh_workbook`.
Fix: Update imports and usage to the current `TableauOperator` and `TableauJobStatusSensor` patterns, which now handle these functionalities directly.
breaking Provider version 3.0.0 requires Apache Airflow 2.2.0 or newer. Installing this provider version on older Airflow instances may automatically upgrade Airflow.
Fix: Ensure your Apache Airflow instance is upgraded to at least version 2.2.0 (or newer for later provider versions) before installing or upgrading to `apache-airflow-providers-tableau` >= 3.0.0.
gotcha Failure to provide the correct `site_id` for Tableau Online or a non-default Tableau Server site can lead to `NotSignedInError: Missing site ID` errors, even if other credentials are correct.
Fix: Always specify the `site_id` parameter in your Tableau connection or operator if you are not using the default Tableau Server site. For the default site, use an empty string (e.g., `site_id=''`).
gotcha Using self-signed SSL certificates can result in `SSLCertVerificationError` even if `openssl s_client` validates the certificate, indicating potential deeper issues with how `tableauserverclient` handles custom CA bundles within Airflow.
Fix: Ensure the certificate chain is correctly configured and accessible by the Airflow worker. If possible, use publicly recognized certificates. For self-signed certificates, ensure the full chain is provided and consider alternative verification methods if the issue persists.

Install

pip install apache-airflow-providers-tableau Install latest version

Imports

TableauOperator

from airflow.providers.tableau.operators.tableau import TableauOperator

TableauHook

from airflow.providers.tableau.hooks.tableau import TableauHook

TableauJobStatusSensor

from airflow.providers.tableau.sensors.tableau_job_status import TableauJobStatusSensor

Quickstart

This example DAG demonstrates how to use the `TableauOperator` to refresh a Tableau workbook. It shows a blocking refresh, meaning the task will wait until the refresh operation on Tableau Server/Online is completed. The `tableau_conn_id` parameter references an Airflow connection configured with Tableau credentials. The `site_id` should be provided if not using the default Tableau site, or an empty string ('') for the default site. Workbook ID can be passed via DAG params or environment variables.

import os
from datetime import datetime

from airflow import DAG
from airflow.providers.tableau.operators.tableau import TableauOperator

with DAG(
    dag_id='example_tableau_refresh_workbook',
    start_date=datetime(2023, 1, 1),
    schedule_interval=None,
    catchup=False,
    tags=['tableau', 'example'],
    params={'workbook_id': os.environ.get('AIRFLOW_WORKBOOK_ID', 'your_workbook_id')}
) as dag:
    refresh_workbook_blocking = TableauOperator(
        task_id='refresh_tableau_workbook_blocking',
        resource='workbooks',
        method='refresh',
        find="{{ params.workbook_id }}",
        site_id=os.environ.get('AIRFLOW_TABLEAU_SITE_ID', ''), # Use '' for default site
        blocking_refresh=True, # Waits until refresh is complete
        tableau_conn_id=os.environ.get('AIRFLOW_TABLEAU_CONN_ID', 'tableau_default')
    )

view raw JSON →