mzecov opened a new issue, #33918: URL: https://github.com/apache/airflow/issues/33918
### Apache Airflow version 2.7.0 ### What happened Variable `{{ prev_start_date_success }}` returns unexpected results when manually triggering DAG. With following DAG definition file ```py with DAG( "demo", schedule="*/1 * * * *", start_date=datetime(2023, 1, 1), max_active_runs=1, catchup=False, ) as dag: t1 = BashOperator( task_id="templated", depends_on_past=False, bash_command="""echo "{{ prev_start_date_success }}" """, ) t1 ``` (one task only displaying `prev_start_date_success`, every minute for demo purposes) With following DAG runs: - DAG run # 1: automatically triggered at 2023-08-30, 12:03:00 UTC; data_interval_start is 2023-08-30, 12:02:00 UTC - DAG run # 2: **manually** triggered **at 2023-08-30, 12:03:34 UTC**; data_interval_start is 2023-08-30, 12:02:00 UTC - DAG run # 3: automatically triggered at 2023-08-30, 12:04:00 UTC; data_interval_start is 2023-08-30, 12:03:00 UTC - DAG run # 4: automatically triggered at 2023-08-30, 12:05:00 UTC; data_interval_start is 2023-08-30, 12:04:00 UTC Logs: - DAG run # 2: `['/bin/bash', '-c', 'echo "2023-08-30T12:03:00.378342+00:00" ']` - DAG run # 3: `['/bin/bash', '-c', 'echo "2023-08-30T12:03:00.378342+00:00" ']` - DAG run # 4: `['/bin/bash', '-c', 'echo "2023-08-30T12:03:34.363933+00:00" ']` ### What you think should happen instead [Here](https://airflow.apache.org/docs/apache-airflow/stable/templates-ref.html): `prev_start_date_success` is *"Start date from prior successful DagRun (if available)."*. But: - DAG run # 3 prints start date of DAG # 1, not # 2 (manually triggered). I thought maybe runs are sorted on `data_interval_start` to get *"prior"*. And this value is the same for DAG run # 1 and # 2. Or maybe manual triggers are excluded. - But why DAG # 4 prints start date of DAG # 2 ? *"prior"* DAG run should be # 3 not # 2 - Is it a bug? - Else, can we explain this in documentation? Note that with `CronTriggerTimetable`, results are different and more coherent (# 3 uses # 2 and # 4 uses # 3). ### How to reproduce - Use upper DAG definition file - Trigger DAG in Airflow GUI (manually), between two schedules - Observe logs of following two DAG runs. ### Operating System Runs in container (image 2.7.0-python3.10) ### Versions of Apache Airflow Providers _No response_ ### Deployment Docker-Compose ### 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 - [X] I agree to follow this project's [Code of Conduct](https://github.com/apache/airflow/blob/main/CODE_OF_CONDUCT.md) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@airflow.apache.org.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org