Redeployment of API server without the scheduler causes dag failure

[connorwang]

Official Helm Chart version

1.18.0 (latest released)

Apache Airflow version

3.1.5

Kubernetes Version

1.33

Helm Chart configuration

No response

Docker Image customizations

No response

What happened

No response

What you think should happen instead

No response

How to reproduce

In my case,

1. Updating apiServer values
2. Upgrade helm install

Anything else

Sometime updating helm values will trigger redeployment of airflow-api-server but not the airflow-scheduler which causes dag failures. The airflow-scheduler deployment has the following error:

airflow.sdk.api.client.ServerResponseError: Invalid auth token: Signature verification failed

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

[boring-cyborg]

Thanks for opening your first issue here! Be sure to follow the issue template! If you are willing to raise PR to address this issue please do so, no need to wait for approval.

[arjav1528]

@shahar1 @potiuk Hi, I would like to work on this issue, could you please assign me

[goldigd05]

Thanks for reporting this issue.

It seems that updating apiServer-related values triggers a redeployment
of the airflow-api-server, but not the airflow-scheduler. This can leave
the scheduler using an outdated auth token, causing DAG failures.

I suspect this may be related to missing apiServer-related config or
secret checksums in the airflow-scheduler deployment template.

I’m happy to investigate this further and submit a PR if this approach
sounds reasonable.