ywang271828 commented on issue #8792:
URL: https://github.com/apache/airflow/issues/8792#issuecomment-2237665807
Another usage case of pushing results to XCOM at failure is to provide
diagnostics that subsequent tasks can analyze and provide a graceful handling.
--
This is an automated me
dstandish commented on issue #8792:
URL: https://github.com/apache/airflow/issues/8792#issuecomment-2237681309
I don't see any harm in just always pushing e.g. by using try / finally.
there's no harm in pushing the xcom if there is something there. downstream
tasks don't need to do anythi
hussein-awala commented on issue #8792:
URL: https://github.com/apache/airflow/issues/8792#issuecomment-1915799699
I'm interested in this proposed feature (btw it is not a bug). I tried to
create something generic as @potiuk suggested, but it's too complicated since
we don't return any resu
hussein-awala commented on issue #8792:
URL: https://github.com/apache/airflow/issues/8792#issuecomment-1915863325
As a workaround, we can use the new callbacks class:
```python
class XComCallbacks(KubernetesPodOperatorCallback):
@staticmethod
def on_pod_completion(*, pod:
marlena-hammond commented on issue #8792:
URL: https://github.com/apache/airflow/issues/8792#issuecomment-1932522623
I am interested in this feature. For, my usecase, it is very helpful being
able to retry the task/DAG based on the exception that is returned.
--
This is an automated messa