-
Notifications
You must be signed in to change notification settings - Fork 16.5k
Open
Labels
area:providerskind:bugThis is a clearly a bugThis is a clearly a bugprovider:cncf-kubernetesKubernetes (k8s) provider related issuesKubernetes (k8s) provider related issues
Description
Apache Airflow Provider(s)
cncf-kubernetes
Versions of Apache Airflow Providers
10.0.1
Apache Airflow version
2.10.5
Operating System
Debian GNU/Linux 12 (bookworm)
Deployment
Astronomer
Deployment details
No response
What happened
The executor_config dictionary used to configure worker pods with the Kubernetes executor does not raise an error if any keys besides "pod_override" or "pod_template_file" are used. Even with invalid keys, the DAG is imported and will run as expected. The invalid executor_config will simply be disregarded.
What you think should happen instead
If any keys besides "pod_override" or "pod_template_file" are used in executor_config, and error (maybe a DAG import error) should be raised.
This should give the user feedback that their executor_config keys are invalid and need to be removed or changed.
How to reproduce
- Create an Airflow instance that uses the Kubernetes Executor.
- Create a DAG with the following code including an invalid key in
executor_config.
import datetime
from airflow.decorators import dag
from airflow.decorators import task
executor_config = {
"key": "value"
}
@dag(start_date=datetime.datetime(2024, 10, 1), schedule="@daily", catchup=False)
def dag_1():
@task(executor_config=executor_config)
def task_1():
print("task 1")
task_1()
dag_1()
- Observe that the DAG will be imported and run on the Airflow instance without any issues.
Anything else
No response
Are you willing to submit PR?
- Yes I am willing to submit a PR!
Code of Conduct
- I agree to follow this project's Code of Conduct
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
area:providerskind:bugThis is a clearly a bugThis is a clearly a bugprovider:cncf-kubernetesKubernetes (k8s) provider related issuesKubernetes (k8s) provider related issues