o-nikolas commented on code in PR #29511: URL: https://github.com/apache/airflow/pull/29511#discussion_r1106346461
########## docs/apache-airflow/ui.rst: ########## @@ -160,6 +160,22 @@ DAG over many runs. ------------ +Landing Times +............. +Airflow landing times are calculated from the task's scheduled time to +the time the task finishes, either with success or another state (see +:ref:`concepts:task-instances`). When subtracting the task duration +(above) from the landing time you will find the adoption time. If the +adoption time exceeds the +:ref:`task adoption timeout <config:celery__task_adoption_timeout>` +then the task will be terminated and, if applicable, rescheduled. Review Comment: Yeah, this area is a real hodgepodge. I think landing times could be and are relevant to other executors (there is surely some delay between scheduling and the task actually running). But it's pronounced in Celery because of the queuing nature. But Celery isn't the only queuing type executor we'll likely have in the future or that people will build as a third party (think AWS Batch). So perhaps it makes more sense to move the logic for those metrics from `celery_executor.py` to `base_executor.py` at some point in time, and then the docs as well. But for the moment, my 2c is that this doc contribution is useful to at least get down on paper somewhere and we can optimize the location in the future. I wouldn't block this PR for that. -- 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]
