Overview

Converts Prefect @task functions that use the tableau-server-client (TSC) Python library to refresh Tableau workbooks or datasource extracts into Orchestra pipeline tasks using the TABLEAU_CLOUD integration. Authentication (Personal Access Token + site ID) moves from environment variables inside the task to a named Orchestra Tableau Cloud connection.

Parameter Mapping

Orchestra YAML Structure

integration: TABLEAU_CLOUD
integration_job: TABLEAU_REFRESH_WORKBOOK   # or TABLEAU_REFRESH_EXTRACT
name: refresh_sales_dashboard
connection: tableau_cloud_prod_12345
parameters:
  project_name: Sales
  workbook_name: Sales Dashboard
depends_on: []
condition: null
tags: []

For a datasource extract refresh:

integration: TABLEAU_CLOUD
integration_job: TABLEAU_REFRESH_EXTRACT
name: refresh_orders_extract
connection: tableau_cloud_prod_12345
parameters:
  project_name: Sales
  datasource_name: Orders Extract
depends_on: []
condition: null
tags: []

Conversion Steps

• Identify whether the task calls server.workbooks.refresh() (→ TABLEAU_REFRESH_WORKBOOK) or server.datasources.refresh() (→ TABLEAU_REFRESH_EXTRACT)
• Extract the workbook/datasource name from the .filter(name=...) call or surrounding code comments
• Extract or infer the Tableau project name (check filter chain, variable names, or ask the user)
• Create an Orchestra Tableau Cloud connection via Integrations → New Connection → Tableau Cloud (server URL, site ID, and PAT name + value — or Connected App client ID, secret ID, secret value, and username if the flow used TSC.JWTAuth)
• Note the connection name assigned by Orchestra (e.g. tableau_cloud_prod_12345)
• Write the Orchestra task YAML using the structure above
• Delete the Prefect @task function and any TSC imports no longer needed
• If the task was called inside a @flow, replace the call with a depends_on: reference in the Orchestra pipeline YAML

Before / After Example

Prefect (before)

import tableau_server_client as TSC
import os
from prefect import task, flow

@task
def refresh_tableau_workbook():
    tableau_auth = TSC.PersonalAccessTokenAuth(
        token_name=os.environ["TABLEAU_TOKEN_NAME"],
        personal_access_token=os.environ["TABLEAU_TOKEN"],
        site_id="mysite",
    )
    server = TSC.Server("https://10ax.online.tableau.com")
    with server.auth.sign_in(tableau_auth):
        workbook = server.workbooks.filter(name="Sales Dashboard").pop()
        server.workbooks.refresh(workbook)

@flow
def tableau_flow():
    refresh_tableau_workbook()

Orchestra YAML (after)

pipeline:
  stage-refresh:
    tasks:
      refresh-sales-dashboard:
        integration: TABLEAU_CLOUD
        integration_job: TABLEAU_REFRESH_WORKBOOK
        name: refresh_sales_dashboard
        connection: tableau_cloud_prod_12345
        parameters:
          project_name: Sales
          workbook_name: Sales Dashboard
        depends_on: []
        condition: null
        tags: []

Gotchas

• project_name is REQUIRED in Orchestra alongside workbook_name — extract it from filter logic, Tableau UI, or ask the user; the task will fail without it
• Site ID and credentials belong on the Orchestra Tableau Cloud connection, not in YAML or env vars
• Auth: PAT or Connected App: the Orchestra Tableau connection takes tableau_auth_type PAT (token name + value) or CONNECTED_APP (client ID, secret ID, secret value, username). A PAT allows only one in-flight request at a time, so prefer a Connected App when several tasks refresh concurrently through one connection. Connected Apps need Tableau Cloud (Dec 2025 release+) or Tableau Server 2025.3+.
• For datasource refresh: use integration_job: TABLEAU_REFRESH_EXTRACT and parameters.datasource_name (not workbook_name)
• There is no native Prefect Tableau block — Tableau tasks in Prefect are always a Python @task wrapping TSC; do not look for a block reference
• The TSC .filter() call returns a generator — .pop() picks the first match; confirm the workbook name is unique in the project before migrating
• server.workbooks.refresh() is async on Tableau Server; Orchestra polls for completion automatically

References

• https://docs.getorchestra.io/docs/integrations/tableau_cloud
• https://tableau.github.io/server-client-python/docs/
• See prefect-alerts-to-orchestra for all notification patterns.

Adding Alerts

See prefect-alerts-to-orchestra for all notification patterns.