Task exited with return code 1 with no errors

[komalrajputtrootech]

Hi all,

Issue

Airflow Version : 2.10.3

I’m running a task in an Airflow DAG that calls a Django function. To initialize Django, I’m using django.setup().

The task consistently fails with exit code 1, but no traceback or error logs are shown.

What I’ve Observed

• Logs before django.setup() are printed correctly.
• Logs after django.setup() never appear.
• The task exits with return code 1 but without any error or stack trace in Airflow logs.

Here’s the log snippet:

[2025-08-25, 07:35:16 UTC] {operators.py:66} INFO - --------------Initializing django ------------------------------------------
[2025-08-25, 07:35:17 UTC] {local_task_job_runner.py:266} INFO - Task exited with return code 1

What I Tried

• Enabled PYTHONFAULTHANDLER → still no traceback.
• Wrapped django.setup() in try/except with traceback.print_exc() → nothing is printed.
• Adjusted Airflow logging to DEBUG → no additional output related to the crash.
• Verified task environment paths and settings are correct.

Ask

Has anyone encountered a situation where django.setup() fails silently inside an MWAA/Airflow task?

• How can I capture the real traceback?
• Could this be related to how MWAA handles task logging/subprocesses?

ip-10-192-21-19.us-east-2.compute.internal
*** Reading remote log from Cloudwatch log_group: airflow-portal-dags-newvpc-Task log_stream: dag_id=house_armed_services_committee_scraper/run_id=manual__2025-08-26T11_37_53.688039+00_00/task_id=scrap_house_armed_services_committee_data/attempt=1.log.
[2025-08-26, 11:37:57 UTC] {local_task_job_runner.py:123} ▼ Pre task execution logs
[2025-08-26, 11:37:57 UTC] {taskinstance.py:2613} INFO - Dependencies all met for dep_context=non-requeueable deps ti=<TaskInstance: house_armed_services_committee_scraper.scrap_house_armed_services_committee_data manual__2025-08-26T11:37:53.688039+00:00 [queued]>
[2025-08-26, 11:37:57 UTC] {taskinstance.py:2613} INFO - Dependencies all met for dep_context=requeueable deps ti=<TaskInstance: house_armed_services_committee_scraper.scrap_house_armed_services_committee_data manual__2025-08-26T11:37:53.688039+00:00 [queued]>
[2025-08-26, 11:37:57 UTC] {taskinstance.py:2866} INFO - Starting attempt 1 of 1
[2025-08-26, 11:37:57 UTC] {taskinstance.py:2889} INFO - Executing <Task(DjangoOperator): scrap_house_armed_services_committee_data> on 2025-08-26 11:37:53.688039+00:00
[2025-08-26, 11:37:57 UTC] {standard_task_runner.py:72} INFO - Started process 2905 to run task
[2025-08-26, 11:37:57 UTC] {standard_task_runner.py:104} INFO - Running: ['airflow', 'tasks', 'run', 'house_armed_services_committee_scraper', 'scrap_house_armed_services_committee_data', 'manual__2025-08-26T11:37:53.688039+00:00', '--job-id', '396', '--raw', '--subdir', 'DAGS_FOLDER/hsas_dag.py', '--cfg-path', '/tmp/tmpwep4g1sf']
[2025-08-26, 11:37:57 UTC] {standard_task_runner.py:105} INFO - Job 396: Subtask scrap_house_armed_services_committee_data
[2025-08-26, 11:37:57 UTC] {task_command.py:467} INFO - Running <TaskInstance: house_armed_services_committee_scraper.scrap_house_armed_services_committee_data manual__2025-08-26T11:37:53.688039+00:00 [running]> on host ip-10-192-21-19.us-east-2.compute.internal
[2025-08-26, 11:37:57 UTC] {taskinstance.py:3132} INFO - Exporting env vars: AIRFLOW_CTX_DAG_OWNER='airflow' AIRFLOW_CTX_DAG_ID='house_armed_services_committee_scraper' AIRFLOW_CTX_TASK_ID='scrap_house_armed_services_committee_data' AIRFLOW_CTX_EXECUTION_DATE='2025-08-26T11:37:53.688039+00:00' AIRFLOW_CTX_TRY_NUMBER='1' AIRFLOW_CTX_DAG_RUN_ID='manual__2025-08-26T11:37:53.688039+00:00'
[2025-08-26, 11:37:57 UTC] {operators.py:59} INFO - ******************************************************************************************
[2025-08-26, 11:37:57 UTC] {operators.py:64} INFO - ['/usr/local/airflow/.local/bin', '/python', '/usr/local/lib/python311.zip', '/usr/local/lib/python3.11', '/usr/local/lib/python3.11/lib-dynload', '/usr/local/airflow/.local/lib/python3.11/site-packages', '/usr/local/lib/python3.11/site-packages', '/usr/local/airflow/config', '/usr/local/airflow/plugins', '/usr/local/airflow/dags', '/usr/local/airflow/.local/lib/python3.11/site-packages/electo_premium_mwaa']
[2025-08-26, 11:37:57 UTC] {operators.py:65} INFO - {'filters': [], 'name': 'operators', 'level': 0, 'parent': <RootLogger root (INFO)>, 'propagate': True, 'handlers': [], 'disabled': False, '_cache': {20: True}, 'manager': <logging.Manager object at 0x7f5253fd3dd0>}
[2025-08-26, 11:37:57 UTC] {operators.py:67} INFO - electo_premium.settings.production
[2025-08-26, 11:37:57 UTC] {operators.py:68} INFO - --------------Initializing django ------------------------------------------
[2025-08-26, 11:39:58 UTC] {local_task_job_runner.py:266} INFO - Task exited with return code 1

[iddqd1]

Add some timeout after django.setup() ...

https://github.com/parisk/airflow-django/blob/master/airflow_django/operators.py
Try to create something similar to this if you want to really use Django from airflow (I do not recommend). Airflow does not really support Django way of things. I love django but Airflow async way of working and Django ...

[komalrajputtrootech]

Any idea what is the use of DJANGO_ENVIRONMENT_HAS_BEEN_SETUP variable. Is it set anywhere else other than operators code ?

Strangely I was able to execute the DAG by replacing django.setup() with apps.populate(settings.INSTALLED_APPS) (which is part of django.setup()). I still don't understand how this worked.

I have done somewhat similar to that in github repo that you shared:

class DjangoOperator(PythonOperator):
    """
    An Airflow Operator for running any Python function that requires the Django environment.

    This operator ensures:
    - Django project path is added to `sys.path`.
    - Django settings module is loaded.
    - `django.setup()` is called before running the task.
    - Allows specifying the callable either directly as a Python function,
      or via a dotted string path (e.g. `"myapp.tasks.my_function"`).
    """

    def __init__(
        self,
        *,
        python_callable: Callable | str,
        op_args: Collection[Any] | None = None,
        op_kwargs: Mapping[str, Any] | None = None,
        templates_dict: dict[str, Any] | None = None,
        templates_exts: Sequence[str] | None = None,
        show_return_value_in_logs: bool = True,
        **kwargs,
    ) -> None:
        # Bypass parent's python_callable check
        self._python_callable = python_callable

        super().__init__(
            python_callable=lambda: None,  # Bypass the AirflowException check
            op_args=op_args,
            op_kwargs=op_kwargs,
            templates_dict=templates_dict,
            templates_exts=templates_exts,
            show_return_value_in_logs=show_return_value_in_logs,
            **kwargs,
        )

    @staticmethod
    def setup_django():
        """
        Sets up Django so that the ORM and other Django-specific components can be used
        within Airflow tasks.

        - Adds the Django project root to `sys.path`.
        - Configures the `DJANGO_SETTINGS_MODULE` environment variable.
        - Calls `django.setup()` to initialize Django.
        """
        # Absolute path to the Django project root where `manage.py` is located
        django_root = "path_to_project"

        if django_root not in sys.path:
            sys.path.insert(0, django_root)

        os.environ.setdefault("DJANGO_SETTINGS_MODULE", "project.settings")

        # Initialize the Django application
        django.setup()

    def resolve_callable(self):
        """
        Resolves the python_callable into a function object.

        Returns
        -------
        callable
            The resolved Python function.

        Raises
        ------
        TypeError
            If python_callable is neither a function nor a dotted string path.
        """
        if callable(self._python_callable):
            return self._python_callable
        elif isinstance(self._python_callable, str):
            module_path, func_name = self._python_callable.rsplit(".", 1)
            module = importlib.import_module(module_path)
            return getattr(module, func_name)
        else:
            raise TypeError(
                "`python_callable` must be a callable or dotted string path"
            )

    def pre_execute(self, context):
        """
        Prepares Django and resolves the Python callable before the task runs.
        """
        self.setup_django()
        self.python_callable = self.resolve_callable()