jstern edited a comment on issue #7133: [AIRFLOW-6535] first pass at exception to fail without retry URL: https://github.com/apache/airflow/pull/7133#issuecomment-573155093 One other question: am I perhaps overcomplicating this? It occurs to me that if a task encounters an AirflowSkipException, we don't bother checking on its retry status or adjusting our messaging based on that: ```python except AirflowSkipException as e: # Recording SKIP # log only if exception has any arguments to prevent log flooding if e.args: self.log.info(e) self.refresh_from_db(lock_for_update=True) self.state = State.SKIPPED self.log.info( 'Marking task as SKIPPED.' 'dag_id=%s, task_id=%s, execution_date=%s, start_date=%s, end_date=%s', self.dag_id, self.task_id, self.execution_date.strftime('%Y%m%dT%H%M%S') if hasattr( self, 'execution_date') and self.execution_date else '', self.start_date.strftime('%Y%m%dT%H%M%S') if hasattr( self, 'start_date') and self.start_date else '', self.end_date.strftime('%Y%m%dT%H%M%S') if hasattr( self, 'end_date') and self.end_date else '') ``` Maybe instead of messing with handle_failure I should just emulate that but with different messaging? Might be cleaner and easier to follow... EDIT: but then I would still need to make sure failure email/callbacks happen ... so I guess I can either leave this structured the way I have it or include those in the except block.
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
