[
https://issues.apache.org/jira/browse/AIRFLOW-4737?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16860763#comment-16860763
]
ASF subversion and git services commented on AIRFLOW-4737:
----------------------------------------------------------
Commit 5db23c6b3d7f5d74bccc0d6e5e35d6c11f67c551 in airflow's branch
refs/heads/master from Victor Villas
[ https://gitbox.apache.org/repos/asf?p=airflow.git;h=5db23c6 ]
[AIRFLOW-4737] Increase and document celery queue name limit (#5383)
> Column queue on task_instance table limit of 50 chars is too low
> ----------------------------------------------------------------
>
> Key: AIRFLOW-4737
> URL: https://issues.apache.org/jira/browse/AIRFLOW-4737
> Project: Apache Airflow
> Issue Type: Improvement
> Components: models
> Affects Versions: 1.10.3
> Reporter: Victor Villas Bôas Chaves
> Assignee: Victor Villas Bôas Chaves
> Priority: Minor
>
> I tried to use an SQS queue named
> "airtest-AirflowStack-FW7GGJK59W9-Tasks-BL0L2W6F5OKF" as message broker, but
> the Scheduler fails to work because the queue name is 51 characters long.
> Why this huge queue name? SQS queue names are automatically generated by
> CloudFormation, and it's best practice to let it do so. Also, when using
> nested stacks, the inner stack also has its name automatically generated and
> appended to the master stack.
> So in short, my master stack is called "airtest" and the queue resource
> "Tasks" (which I believe are quite reasonably sized), but that huge name is
> what you end up in the end. Therefore the use case for longer queue names is
> there.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
