vincbeck commented on code in PR #28502:
URL: https://github.com/apache/airflow/pull/28502#discussion_r1062612567
##########
airflow/dag_processing/processor.py:
##########
@@ -365,8 +365,10 @@ def __init__(self, dag_ids: list[str] | None,
dag_directory: str, log: logging.L
self._dag_directory = dag_directory
self.dag_warnings: set[tuple[str, str]] = set()
+ @staticmethod
+ @internal_api_call
@provide_session
- def manage_slas(self, dag: DAG, session: Session = None) -> None:
+ def manage_slas(dag_folder, dag_id: str, log: logging.Logger, session:
Session = NEW_SESSION) -> None:
Review Comment:
I agree, it feels also very wrong to me but maybe with some context it will
help understanding it :).
All methods decorated with `@internal_api_call` are meant to be called
either directly (like today) or through a new component: internal API ([see PR
here](https://github.com/apache/airflow/pull/27892)). This decision is based on
the the flag `database_access_isolation`. What we want to achieve here is when
the methods are called directly, the behavior should be the same as today,
hence the logger should be the same. However, when these methods are called
through the internal API, we want the logs to be separated from the current
execution. Then, in that case we want a different logger. This new logger is
created by the internal API itself, [see PR
here](https://github.com/apache/airflow/pull/28476).
I hope it helps to understand and feel free to give me your thoughts given
this context
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
