kimyen opened a new pull request, #36078:
URL: https://github.com/apache/airflow/pull/36078
<!--
Licensed to the Apache Software Foundation (ASF) under one
or more contributor license agreements. See the NOTICE file
distributed with this work for additional information
regarding copyright ownership. The ASF licenses this file
to you under the Apache License, Version 2.0 (the
"License"); you may not use this file except in compliance
with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing,
software distributed under the License is distributed on an
"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
KIND, either express or implied. See the License for the
specific language governing permissions and limitations
under the License.
-->
<!--
Thank you for contributing! Please make sure that your code changes
are covered with tests. And in case of new features or big changes
remember to adjust the documentation.
Feel free to ping committers for the review!
In case of an existing issue, reference it using one of the following:
closes: #ISSUE
related: #ISSUE
How to write a good git commit message:
http://chris.beams.io/posts/git-commit/
-->
### Use case
We use Airflow to build our data pipelines. Our Airflow DAG graph includes
600+ DAGs, highly dependent on one another. A DAG may include ingestion for
multiple tables from the same source. A downstream DAG may only need data from
a single table and only need to take dependency on a single task instead of an
entire DAG.
Occasionally when an upstream DAG timeout due to whatever reason (higher
volume of data, infrastructure issues, ...) we have an on-call rotation that
fixes the issues based on the severity of the DAG. In general, we value no data
over wrong/missing data. Hence, when an upstream DAG timeout, we would address
the issue and ensure the upstream DAG rerun successfully prior to re-enable
(clear) all the downstream DAGs.
With the default settings, when the DAG timeout, the unfinished tasks are
marked as `skipped`, which is a successful state. With this, downstream DAG
whose depends on the unfinished task will start even though the data not yet
arrived/ not completed.
We previously fork `apache/airflow` to change this behavior to mark
unfinished tasks as `failed`. We are working on getting rid of our fork and
hope that we can expose this option to modify the state of unfinished tasks for
our use case.
### Other considerations
Alternatively, with dataset, we can eventually convert our (600+) DAGs to
depend on the data arriving rather than the upstream task instance. However, we
are currently running Airflow 2.0.2 and working on upgrading to Airflow 2.7.
Making our DAG depends on dataset is not yet available for our users. This
change will enable us upgrade directly to `apache/airflow` v2.7 without
maintaining our fork nor impacting our users.
### Summary of changes
This pull request introduces a new configuration parameter
`default_unfinished_ti_state` under `core` section. This parameter is used to
define the state of unfinished tasks when a DAG times out. The changes involve
adding the new configuration parameter in the `config.yml` file and modifying
the `_schedule_dag_run` method in `scheduler_job_runner.py` to use this new
parameter.
Configuration changes:
* <a
href="diffhunk://#diff-0a480a8b563e200f749424d7e761555e543588ff39a0b11cce7e34d522f38e0eR258-R264">`airflow/config_templates/config.yml`</a>:
Introduced a new configuration parameter `default_unfinished_ti_state` under
the `core` section. This parameter is used to specify the state of unfinished
tasks when a DAG times out. The default value for this parameter is set to
"skipped".
Code changes:
* <a
href="diffhunk://#diff-b0491913f69327937706aea8fc77a71efeb979897898e405ade2b162ad862476R1461-R1463">`airflow/jobs/scheduler_job_runner.py`</a>:
In the `_schedule_dag_run` method, the state of unfinished tasks is now set to
the value of the new configuration parameter `default_unfinished_ti_state`
instead of being hard-coded to `TaskInstanceState.SKIPPED`.
<!-- Please keep an empty line above the dashes. -->
---
**^ Add meaningful description above**
Read the **[Pull Request
Guidelines](https://github.com/apache/airflow/blob/main/CONTRIBUTING.rst#pull-request-guidelines)**
for more information.
In case of fundamental code changes, an Airflow Improvement Proposal
([AIP](https://cwiki.apache.org/confluence/display/AIRFLOW/Airflow+Improvement+Proposals))
is needed.
In case of a new dependency, check compliance with the [ASF 3rd Party
License Policy](https://www.apache.org/legal/resolved.html#category-x).
In case of backwards incompatible changes please leave a note in a
newsfragment file, named `{pr_number}.significant.rst` or
`{issue_number}.significant.rst`, in
[newsfragments](https://github.com/apache/airflow/tree/main/newsfragments).
--
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]
