databricks sdk version installed in serverless compute differ

[Chinu]

Hello, 

I've encountered an issue with my Python notebook where app.list() is failing in some serverless compute clusters but works fine in others. After investigating further, I noticed the following version differences:

• Working Cluster: SDK version 0.40.0

• Failing Cluster: SDK version 0.20.0

This version discrepancy might be the root cause, though I couldn't find explicit documentation confirming this.

My Questions:

1. How can I definitively verify if the SDK version is causing this issue?

2. What's the recommended way to update the SDK version in serverless compute? (I assumed serverless would use the same SDK version across all clusters since it's managed by Databricks.)

Any guidance would be appreciated.

[Chinu]

oh, I found the doc and confirmed that list() method is added with version 0.27.0

• Added list() method for w.apps workspace-level service.

https://github.com/databricks/databricks-sdk-py/blob/v0.40.0/CHANGELOG.md

Now, how can I update this to newer version?

[Advika]

Hello @Chinu!

Since you're using serverless compute, upgrading or specifying a Python package like databricks-sdk is recommended through the notebook’s Environment side panel.
You can either update to the latest environment version or manually add the desired package as a dependency.

You can refer to this documentation: Configure the serverless environment