如何使用 Python 在 Airflow 中的另一个 DAG 成功时触发 DAG？

Question

我有一个 python DAGParent Job和 DAG Child Job。中的任务Child Job应该在成功完成Parent Job每天运行的任务时触发。如何添加外部作业触发器？

我的代码

from datetime import datetime, timedelta
from airflow import DAG
from airflow.operators.postgres_operator import PostgresOperator
from utils import FAILURE_EMAILS

yesterday = datetime.combine(datetime.today() - timedelta(1), datetime.min.time())


default_args = {
    'owner': 'airflow',
    'depends_on_past': False,
    'start_date': yesterday,
    'email': FAILURE_EMAILS,
    'email_on_failure': False,
    'email_on_retry': False,
    'retries': 1,
    'retry_delay': timedelta(minutes=5)
}

dag = DAG('Child Job', default_args=default_args, schedule_interval='@daily')

execute_notebook = PostgresOperator(
  task_id='data_sql',
  postgres_conn_id='REDSHIFT_CONN',
  sql="SELECT * FROM athena_rs.shipments limit 5",
  dag=dag
)

Answer 1

答案已经在这个线程中了。下面是演示代码：

父母亲：

from datetime import datetime
from airflow import DAG
from airflow.operators.dummy_operator import DummyOperator

default_args = {
    'owner': 'airflow',
    'depends_on_past': False,
    'start_date': datetime(2020, 4, 29),
}

dag = DAG('Parent_dag', default_args=default_args, schedule_interval='@daily')

leave_work = DummyOperator(
    task_id='leave_work',
    dag=dag,
)
cook_dinner = DummyOperator(
    task_id='cook_dinner',
    dag=dag,
)

leave_work >> cook_dinner

儿童dag：

from datetime import datetime, timedelta
from airflow import DAG
from airflow.operators.dummy_operator import DummyOperator
from airflow.operators.sensors import ExternalTaskSensor

default_args = {
    'owner': 'airflow',
    'depends_on_past': False,
    'start_date': datetime(2020, 4, 29),
}

dag = DAG('Child_dag', default_args=default_args, schedule_interval='@daily')

# Use ExternalTaskSensor to listen to the Parent_dag and cook_dinner task
# when cook_dinner is finished, Child_dag will be triggered
wait_for_dinner = ExternalTaskSensor(
    task_id='wait_for_dinner',
    external_dag_id='Parent_dag',
    external_task_id='cook_dinner',
    start_date=datetime(2020, 4, 29),
    execution_delta=timedelta(hours=1),
    timeout=3600,
)

have_dinner = DummyOperator(
    task_id='have_dinner',
    dag=dag,
)
play_with_food = DummyOperator(
    task_id='play_with_food',
    dag=dag,
)

wait_for_dinner >> have_dinner
wait_for_dinner >> play_with_food

图片：

达格

Parent_dag

Child_dag

Answer 2

如请求由@pankaj，我在此添加一个片段描绘反应性触发使用TriggerDagRunOperator（而不是基于轮询的触发的ExternalTaskSensor）

from typing import List

from airflow.models.baseoperator import BaseOperator
from airflow.models.dag import DAG
from airflow.operators.dagrun_operator import TriggerDagRunOperator
from airflow.utils.trigger_rule import TriggerRule

# DAG object
my_dag: DAG = DAG(dag_id='my_dag',
                  start_date=..)
..
# a list of 'tail' tasks: tasks that have no downstream tasks
tail_tasks_of_first_dag: List[BaseOperator] = my_magic_function_that_determines_all_tail_tasks(..)
..

# our trigger task
my_trigger_task: TriggerDagRunOperator = TriggerDagRunOperator(dag=my_dag,
                                                               task_id='my_trigger_task',
                                                               trigger_rule=TriggerRule.ALL_SUCCESS,
                                                               external_dag_id='id_of_dag_to_be_triggered')
# our trigger task should run when all 'tail' tasks have completed / succeeded
tail_tasks_of_first_dag >> my_trigger_task

请注意，代码片段仅供参考；它没有经过测试

注意事项/参考资料

• 获取所有气流叶节点/任务
• 将顶级 DAG 连接在一起
• 气流触发规则“all_done”和“all_success”有什么区别？