Continuous timetable causes false dag triggering if dag has already run and start_date is in future #45081
Description
Apache Airflow version
2.10.4
If "Other Airflow 2 version" selected, which one?
No response
What happened?
hello,
here is a sample dag
import pendulum
import airflow
import logging
from airflow import DAG
from datetime import timedelta, datetime
from airflow.utils.dates import days_ago
from airflow.operators.python import PythonOperator
from airflow.decorators import task
local_tz = pendulum.timezone("Europe/Paris")
default_args = {
'owner': 'interfaces',
'depends_on_past': False,
'retries': 0,
'retry_delay': timedelta(minutes=1),
}
with DAG(
dag_id='test_continous',
default_args=default_args,
schedule='@continuous',
start_date=datetime(2024,12,19, 16, 40, tzinfo=local_tz),
end_date= datetime(2024,12,19, 16, 46, tzinfo=local_tz),
max_active_runs=1,
catchup=False,
) as dag:
@task()
def the_task():
import time
time.sleep(25)
the_task()the first time you activate it, it will correctly wait for start_date to trigger dag run but if it has already run, every seconds it will trigger a "void" dag run (no task started just a row inserted in the dag_run table)
What you think should happen instead?
Dag shouldn't be triggered until the new start_date is passed
How to reproduce
use the given dag, change start_date and end_date in order to let dag runs for a few minutes. when end_date is passed change start_date and end_date in order to make dag runs in future.
Operating System
debian 11
Versions of Apache Airflow Providers
No response
Deployment
Official Apache Airflow Helm Chart
Deployment details
No response
Anything else?
No response
Are you willing to submit PR?
- Yes I am willing to submit a PR!
Code of Conduct
- I agree to follow this project's Code of Conduct