[ https://issues.apache.org/jira/browse/AIRFLOW-4482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xiaodong DENG resolved AIRFLOW-4482. ------------------------------------ Resolution: Resolved Fix Version/s: 2.0.0 > Add execution_date to trigger DAG run API response > -------------------------------------------------- > > Key: AIRFLOW-4482 > URL: https://issues.apache.org/jira/browse/AIRFLOW-4482 > Project: Apache Airflow > Issue Type: Improvement > Components: api > Affects Versions: 1.10.3 > Reporter: Pavel Shuvalov > Assignee: Pavel Shuvalov > Priority: Major > Fix For: 2.0.0 > > > We are using experimental REST API for automating Airflow from Jenkins and > our workflow looks like this: > * Jenkins job triggers DAG run (using _POST > /api/experimental/dags/*<dag_id>*/dag_runs_) > * Airflow API returns response like this: > {code:java} > {"message":"Created <DagRun dag_id @ 2019-05-08 11:42:43+00:00: run_id, > externally triggered: True>"} {code} > * Jenkins job parses the response for _*execution_date*_ > (2019-05-08T11:42:43 for response above) > * Jenkins job further uses execution_date to check DAG run state using _POST > /api/experimental/dags/*<dag_id>*/dag_runs/*<execution_date>*_ > * Also Jenkins job generates DAG run UI link > Here is [the > code|https://github.com/doublescoring/jenkins-pipeline-goodness/blob/master/src/main/groovy/airflow.groovy]. > Actually it is not a good idea to parse message string for execution_date. So > it is proposed to add _*execution_date*_ in DAG run trigger API response as a > field. It makes it possible to get execution_date directly from the response > without dirty and unpredictable parsing. After this improvement API response > will be like this: > > {code:java} > {'execution_date': '2019-05-08T07:03:09+00:00', 'message': 'Created <DagRun > dag_id @ 2019-05-08 07:03:09+00:00: manual__2019-05-08T07:03:09+00:00, > externally triggered: True>'} > {code} > > -- This message was sent by Atlassian JIRA (v7.6.3#76005)