DagBundle Import Error Detected but Not Displayed in UI

[carlos54]

Apache Airflow version

Other Airflow 2/3 version (please specify below)

If "Other Airflow 2/3 version" selected, which one?

3.1.4

What happened?

Description :
The Dag Processor correctly identifies and logs an import error within a DAG file, but the corresponding error message and Stack Trace are not visible in the Airflow User Interface (no "import Errors" tab).

DAG File Processing Stats
Bundle File Path PID Current Duration # DAGs # Errors Last Duration Last Run At
-------- ---------------------------------- ----- ------------------ -------- ---------- --------------- -------------------
sandbox sandbox_dag005_dataset_producer.py 1 0 0.26s 2025-12-10T11:38:29
sandbox sandbox_dag_demo.py 0 1 0.21s 2025-12-10T11:38:29

Error in log dag processor log file are correctly catch :

{"timestamp":"2025-12-10T11:33:23.270722Z","level":"error","event":"AirflowException(\"Task Policy validation failed for 'DO_DATASET':\\n[DATASET OUTLETS] Dataset URI 'event_data01_dispo' in
 task 'DO_DATASET' must start with 'sandbox_'.\")","logger":"airflow.models.dagbag.DagBag","filename":"dagbag.py","lineno":556,"error_detail":[{"exc_type":"AirflowException","exc_value":"Tas
k Policy validation failed for 'DO_DATASET':\n[DATASET OUTLETS] Dataset URI 'event_data01_dispo' in task 'DO_DATASET' must start with 'sandbox_'.","exc_notes":[],"syntax_error":null,"is_caus
e":false,"frames":[{"filename":"/usr/local/lib/python3.12/site-packages/airflow/models/dagbag.py","lineno":552,"name":"bag_dag"},{"filename":"/usr/local/lib/python3.12/site-packages/airflow/settings.py","lineno":183,"name":"task_policy"},{"filename":"/usr/local/lib/python3.12/site-packages/pluggy/_hooks.py","lineno":512,"name":"__call__"},{"filename":"/usr/local/lib/python3.12/site-packages/pluggy/_manager.py","lineno":120,"name":"_hookexec"},{"filename":"/usr/local/lib/python3.12/site-packages/pluggy/_callers.py","lineno":167,"name":"_multicall"},{"filename":"/usr/local/lib/python3.12/site-packages/pluggy/_callers.py","lineno":121,"name":"_multicall"},{"filename":"/opt/airflow/config/cluster_policy.py","lineno":173,"name":"task_policy"}],"is_group":false,"exceptions":[]}]}

This prevents developers from quickly identifying the source of the failure, especially those caused by the cluster_policy.

What you think should happen instead?

The "Import Errors" view in the Airflow UI should be display .

How to reproduce

The dag_processor ariflow.cfg:
dag_bundle_config_list: '{{ .Values.DAGS.bundle_list }}'
dag_bundle_storage_path: '/dags/airflow/bundles'
refresh_interval: 60
bundle_refresh_check_interval : 15
print_stats_interval: 60
parsing_processes: 2
dag_file_processor_timeout: 60
min_file_process_interval: 30
stale_dag_threshold : 300

Operating System

redhat/ubi9:9.6

Versions of Apache Airflow Providers

apache-airflow==3.1.4
structlog==25.5.0
psycopg2-binary==2.9.11
asyncpg==0.31.0
apache-airflow-providers-fab==3.0.3
apache-airflow-providers-redis==4.4.0
apache-airflow-providers-git==0.1.0
apache-airflow-providers-cncf-kubernetes==10.11.0
flask-limiter==3.12
redis==5.3.1
authlib==1.6.5
PyJWT==2.10.1
cryptography==42.0.8
requests==2.32.5

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

[bil-ly]

@potiuk @carlos54 I can look into this

[carlos54]

Thanks! Because without UI feedback, a developer cannot know what is wrong with their DAG.

[potiuk]

@potiuk @carlos54 I can look into this

Please @bil-ly don't mention people who might or might not be related to the issue/ PR - if you want to be assigned just mention it in the comment - and someone will assign you - it does not have to be me.

[carlos54]

airflow config list

`
[core]
dags_folder = /opt/airflow/dags
hostname_callable = airflow.utils.net.getfqdn
might_contain_dag_callable = airflow.utils.file.might_contain_dag_via_default_heuristic
default_timezone = Europe/Paris
executor = KubernetesExecutor
auth_manager = airflow.providers.fab.auth_manager.fab_auth_manager.FabAuthManager
simple_auth_manager_users = admin:admin
simple_auth_manager_all_admins = False
parallelism = 5
max_active_tasks_per_dag = 2
dags_are_paused_at_creation = True
max_active_runs_per_dag = 1
max_consecutive_failed_dag_runs_per_dag = 0
load_examples = False
plugins_folder = /opt/airflow/plugins
execute_tasks_new_python_interpreter = False
fernet_key = <FERNET_KEY_MASKED>
dagbag_import_timeout = 30.0
dagbag_import_error_tracebacks = True
dagbag_import_error_traceback_depth = 10
default_impersonation =
security =
unit_test_mode = False
allowed_deserialization_classes = airflow.*
allowed_deserialization_classes_regexp =
killed_task_cleanup_time = 60
dag_run_conf_overrides_params = True
dag_discovery_safe_mode = True
dag_ignore_file_syntax = glob
default_task_retries = 0
default_task_retry_delay = 300
max_task_retry_delay = 86400
default_task_weight_rule = downstream
task_success_overtime = 20
default_task_execution_timeout =
min_serialized_dag_update_interval = 30
compress_serialized_dags = False
min_serialized_dag_fetch_interval = 10
max_num_rendered_ti_fields_per_task = 30
xcom_backend = airflow.sdk.execution_time.xcom.BaseXCom
lazy_load_plugins = false
lazy_discover_providers = True
hide_sensitive_var_conn_fields = True
sensitive_var_conn_names =
default_pool_task_slot_count = 128
max_map_length = 1024
daemon_umask = 0o077
test_connection = Disabled
max_templated_field_length = 4096
execution_api_server_url = http://airflow-api-server:8080/execution/
colored_console_log = False
fernet_key_cmd = cat /vault/secrets/fernet-key
remote_logging = False

[database]
alembic_ini_file_path = alembic.ini
sql_alchemy_conn = postgresql+psycopg2://proxy_user:***@db-proxy.internal:6432/airflow_db
sql_engine_encoding = utf-8
sql_alchemy_pool_enabled = true
sql_alchemy_pool_size = 5
sql_alchemy_max_overflow = 10
sql_alchemy_pool_recycle = 1800
sql_alchemy_pool_pre_ping = true
max_db_retries = 3
check_migrations = True
migration_batch_size = 10000

[logging]
base_log_folder = /opt/airflow/logs
remote_logging = false
logging_level = DEBUG
logging_config_class =
colored_console_log = False
log_format =
simple_log_format = %%(asctime)s %%(levelname)s - %%(message)s
dag_processor_log_target = file
log_formatter_class = airflow.utils.log.timezone_aware.TimezoneAware
min_length_masked_secret = 5
task_log_prefix_template =
log_filename_template = dag_id={{ ti.dag_id }}/run_id={{ ti.run_id }}/task_id={{ ti.task_id }}/{%% if ti.map_index >= 0 %%}map_index={{ ti.map_index }}/{%% endif %%}attempt={{ try_number|default(ti.try_number) }}.log
task_log_reader = task
file_task_handler_new_folder_permissions = 0o775
file_task_handler_new_file_permissions = 0o664

[api]
enable_swagger_ui = True
secret_key = <API_SECRET_MASKED>
expose_config = False
expose_stacktrace = true
base_url = https://airflow.internal
host = 0.0.0.0
port = 8080
workers = 1
worker_timeout = 120
maximum_page_limit = 100
fallback_page_limit = 50
grid_view_sorting_order = topological
log_fetch_timeout_sec = 5
hide_paused_dags_by_default = False
page_size = 50
auto_refresh_interval = 3
secret_key_cmd = cat /vault/secrets/webserver-key

[api_auth]
jwt_expiration_time = 86400
jwt_cli_expiration_time = 3600
jwt_secret = cat /vault/secrets/webserver-key
algorithm = RS256
audience = https://api.internal/api/v2
issuer = https://api-stg.internal
jwks_url = https://auth.internal/sps/jwks

[smtp]
smtp_host = mail-relay.internal
smtp_starttls = true
smtp_ssl = false
smtp_port = 25
smtp_mail_from = noreply@internal
smtp_timeout = 30
smtp_retry_limit = 5
smtp_password = cat /vault/secrets/smtp-password
smtp_user = noreply@internal

[kubernetes]
namespace = airflow-prod
worker_container_repository = registry.internal/airflow
worker_container_tag = 0.0.1
delete_worker_pods = True
multi_namespace_mode = true
in_cluster = True

[webserver]
rbac = True
enable_proxy_fix = True

`

[carlos54]

FYI
The Airflow table in database shows no errors despite the DAG processing failure.:

./logs/dag_processor/latest/sandbox/sandbox_base_test_error.py.log :

{"timestamp":"2026-01-27T14:21:01.795727Z","level":"error","event":"Failed to import: /dags/airflow/bundles/sandbox/tracking_repo/sandbox_base_test_error.py","logger":"airflow.models.dagbag.DagBag","filename":"dagbag.py","lineno":415,"error_detail":[{"exc_type":"SyntaxError","exc_value":"invalid syntax. Perhaps you forgot a comma? (sandbox_base_test_error.py, line 63)","exc_notes":[],"syntax_error":{"offset":25,"filename":"/dags/airflow/bundles/sandbox/tracking_repo/sandbox_base_test_error.py","line":"...

SELECT * FROM import_error; id | timestamp | filename | bundle_name | stacktrace ----+-----------+----------+-------------+------------ (0 rows)

[uplsh580]

Is there any ongoing work on this issue?
I'd like to distribute a fixed version to our organization.
If not, can I fix it myself?
I'm asking cautiously, as there's already an assignee.

[bil-ly]

@uplsh580 you can take a look at it, i hadn't realised(didn't get notification email) , i hadn't done a lot , so you can implement the fix

[uplsh580]

This issue seems to be occurring in LocalDagBundle as well.