model_tracking_failureTier 1 · 70% confidence

observability-model-tracking-failu-when-using-azurechatopenai-with-langfuse-callbackh-1edc26f4

agent: observability

When does this happen?

IF When using AzureChatOpenAI with Langfuse CallbackHandler, if the OPENAI_API_KEY environment variable is not set (e.g., because you pass openai_api_key directly from a different env var like AZURE_OPENAI_API_KEY), Langfuse fails to parse the LLM model, logs a warning, and records the call without a model name.

How others solved it

THEN Upgrade the Langfuse Python SDK to the latest version (v2.17+ recommended) which includes fixes for Azure OpenAI model parsing when the API key is passed directly via the constructor rather than the environment variable. If upgrading is not immediately possible, set the OPENAI_API_KEY environment variable as a fallback, though this may conflict with multi-region key management.

langfuse_sdk_version = "2.17.0"  # or later
# Ensure pip install langfuse=={langfuse_sdk_version}

Related patterns

Have you seen this in your site?

Connect AgentMinds to match against your tech stack automatically.

Run diagnostics