Databricks MLflow traces- spans are missing/empty

[neelamkpatil]

Describe the Bug

When viewing traces in Databricks MLflow, the trace details page does not display any spans (trace breakdown), even though the trace exists. see attached screenshot-

The same agent execution correctly shows all spans in the appfnd CLS instance, suggesting the spans are being generated but are not being displayed or ingested correctly in Databricks.

cloud sdk version used- 0.33.2
Note-This issue has been observed since 20 July 2026. Traces generated before this date were populated correctly. Please refer to the traces from 17 July for comparison.

Steps to Reproduce

1. For agent repo- https://github.tools.sap/Montana-AI-Singapore/currency-converter-agent-appfnd run the bruno collection to invoke the agent via JSON RPC endpoint
2. Track the the generated trace in Databricks MLflow 'traces' dashboard
3. Observe that only the Trace ID is visible and the trace contains no span details.

Expected Behavior

The trace should display all associated spans, similar to what is visible in the local CLS instance.

Screenshots

No response

Used Versions

• Python version via python --version: 3.12
• SAP Cloud SDK for Python version: 0.33.2
• Framework version (if applicable, e.g., Flask, FastAPI): ...

Installed packages via pip list or uv pip list

Package list here

Code Examples

# Your code here

Stack Trace

No response

Log File

Log file

...

Affected Development Phase

Development

Impact

Impaired

Timeline

No response

[neelamkpatil]

Same issue persisted after upgrading sdk to 0.37.0

[neelamkpatil]

Adding observations from Natania, to check if they are relevant for the issue-

Seems like the issue is that there are multiple spans with null parent_span_id for the same trace_id, so the trace was not processed correctly

[paramtatini]

Could you invoke the agent again and verify whether the latest traces are showing all the spans in the Databricks UI?