You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@airflow.apache.org by "Krzysztof Madejski (Jira)" <ji...@apache.org> on 2020/03/18 15:00:00 UTC

[jira] [Created] (AIRFLOW-7083) Improve Concepts -> Trigger Rules documentation

Krzysztof Madejski created AIRFLOW-7083:
-------------------------------------------

             Summary: Improve Concepts -> Trigger Rules documentation
                 Key: AIRFLOW-7083
                 URL: https://issues.apache.org/jira/browse/AIRFLOW-7083
             Project: Apache Airflow
          Issue Type: Improvement
          Components: documentation
    Affects Versions: 1.10.9
            Reporter: Krzysztof Madejski


I find documentation on [https://airflow.apache.org/docs/stable/concepts.html#trigger-rules] quite confusing.

By all_failed is mentioned "all parents are in a failed or upstream_failed state". No other rule has such detailed information.

For example: "one_failed: fires as soon as at least one parent has failed, it does not wait for all parents to be done". Does it mean that at least one parent is in `failed` state? Or is task fired also if one is in `upstream_failed` state?

"Skipped tasks will cascade through trigger rules" is confusing. What does "cascade" exactly mean? How is state propagated?

Also: Is "none_failed" and "all_succeeded" actually testing the same condition? If not, what is the difference.


Finally:
 * the documentation does not mention how is the state of a DAG computed. Is it an all_succeeded rule taking as parents all the leaf nodes of the DAG?
 * documentations uses `state` and `status` words interchangeably. Please decide on one.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)