Troubleshooting OpenInference Auto Traces in Anthropic Python SDK

·Aug 12, 2025 02:52 PM

Howdy! 👋🏼 Working with Arize Phoenix for the first time. I am attempting to test out the openinference auto traces for the Anthropic Python SDK. Following the documentation, I'm still not able to see any traces logged in my space/project. Is this a known issue? I read somewhere that if I'm using beta features (i.e. the MCP server configs), it may not work.

👋1

3 comments

· Sorted by Oldest

RunLLM
APP
·
Auto-tracing for the Anthropic Python SDK is supported via the OpenInference Anthropic Instrumentor, and traces should appear in Phoenix if set up correctly. There are no documented general issues with Anthropic auto-tracing, but if you are using beta features like MCP server configs, there may be compatibility limitations, as MCP tracing is handled differently and may not generate its own telemetry unless both client and server are instrumented for OpenTelemetry traces. Ensure you are using the latest versions and have set up both the instrumentor and Phoenix endpoint properly, as shown in the <https://github.com/Arize-ai/openinference/blob/main/python/instrumentation/openinference-instrumentation-anthropic/README.md|OpenInference Anthropic Instrumentation README> and <https://arize.com/docs/phoenix/integrations/model-context-protocol/mcp-tracing|MCP Tracing docs>. Would you like a step-by-step setup guide or troubleshooting checklist? Sources: - <https://github.com/Arize-ai/openinference/blob/main/python/instrumentation/openinference-instrumentation-anthropic/README.md|OpenInference Anthropic Instrumentation README> - <https://arize.com/docs/phoenix/integrations/model-context-protocol/mcp-tracing|MCP Tracing docs>
💡 Hint: Mention RunLLM in the thread for followups.
Dustin N.
·
Hi Jacob! Can you give us a little bit more detail about how you’re setting up your instrumentation and where you’re looking for traces?
Jacob M.
·
I was able to resolve this. Thank you!