Task instance finished with state failed, but the task instance's state attribute is queued

### Under which category would you file this issue? Providers ### Apache Airflow version 3.2.0 ### What happened and how to reproduce it? From a clean fresh install Airflow v3.2.0, after a manual (or logical date in the future, or scheduled task) task launch, it dies with the message: ``` 2026-04-09T18:46:08.340118Z [info ] Task execute_workload[26ecc9e5-5612-4dc7-8987-3334f991be3b] received [celery.worker.strategy] loc=strategy.py:161 2026-04-09T18:46:08.340637Z [debug ] TaskPool: Apply <function fast_trace_task at 0x7f29a2fd4680> (args:('execute_workload', '26ecc9e5-5612-4dc7-8987-3334f991be3b', {'argsrepr':... kwargs:{}) [celery.pool] loc=base.py:149 2026-04-09T18:46:08.341924Z [debug ] Task accepted: execute_workload[26ecc9e5-5612-4dc7-8987-3334f991be3b] pid:1034973 [celery.worker.request] loc=request.py:525 2026-04-09T18:46:08.347260Z [info ] [26ecc9e5-5612-4dc7-8987-3334f991be3b] Executing workload in Celery: dag_rel_path=PurePosixPath('AplicacionesIT/airflow-log-cleanup.py') bundle_info=BundleInfo(name='dags-folder', version=None) log_path='dag_id=airflow-log-cleanup/run_id=manual__2026-04-09T18:40:35.430296+00:00/task_id=log_cleanup_worker_num_1_dir_0/attempt=2.log' ti=TaskInstanceDTO(id=UUID('019d738b-ce7e-78fa-813d-445c019749f6'), dag_version_id=UUID('019d6e44-9c6d-7396-9815-819c356c2808'), task_id='log_cleanup_worker_num_1_dir_0', dag_id='airflow-log-cleanup', run_id='manual__2026-04-09T18:40:35.430296+00:00', try_number=2, map_index=-1, pool_slots=1, queue='default', priority_weight=1, executor_config=None, parent_context_carrier=None, context_carrier={'traceparent': '00-9d7b223ec75d27ba68df0ea91638ab60-2e2651f84e3e6e05-01'}) sentry_integration='sentry_sdk.integrations.celery.CeleryIntegration' type='ExecuteTask' [airflow.providers.celery.executors.celery_executor_utils] loc=celery_executor_utils.py:193 2026-04-09T18:46:08.349156Z [debug ] Connecting to execution API server [supervisor] loc=supervisor.py:2077 server=https://air3d01xdl.**…