[Bug]: LiteLLM SDK Client does not send out metadata upstream

[emerzon]

What happened?

Using LiteLLM completion client, the metadata field is ignored.

Sample code:

import litellm

litellm.api_base = ""  # Replace with your LiteLLM server URL
litellm.api_key = "sk-xxx"
litellm._turn_on_debug()

# Create a completion
response = litellm.completion(
    model="gpt-4.1-nano",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Write a haiku about the ocean"},
    ],
    metadata={"session_id": "abc-123"}
)

# Print the completion
print(response.choices[0].message["content"])

Gives the following output, without the metadata being forwarded:

18:06:08 - LiteLLM:DEBUG: utils.py:348 - Request to litellm:
18:06:08 - LiteLLM:DEBUG: utils.py:348 - litellm.completion(model='gpt-4.1-nano', messages=[{'role': 'system', 'content': 'You are a helpful assistant.'}, {'role': 'user', 'content': 'Write a haiku about the ocean'}], metadata={'session_id': 'abc-123'})
18:06:08 - LiteLLM:DEBUG: utils.py:348 - 

18:06:08 - LiteLLM:DEBUG: litellm_logging.py:475 - self.optional_params: {}
18:06:08 - LiteLLM:DEBUG: utils.py:348 - SYNC kwargs[caching]: False; litellm.cache: None; kwargs.get('cache')['no-cache']: False
18:06:08 - LiteLLM:INFO: utils.py:3230 - 
LiteLLM completion() model= gpt-4.1-nano; provider = openai
18:06:08 - LiteLLM:DEBUG: utils.py:3233 - 
LiteLLM: Params passed to completion() {'model': 'gpt-4.1-nano', 'functions': None, 'function_call': None, 'temperature': None, 'top_p': None, 'n': None, 'stream': None, 'stream_options': None, 'stop': None, 'max_tokens': None, 'max_completion_tokens': None, 'modalities': None, 'prediction': None, 'audio': None, 'presence_penalty': None, 'frequency_penalty': None, 'logit_bias': None, 'user': None, 'custom_llm_provider': 'openai', 'response_format': None, 'seed': None, 'tools': None, 'tool_choice': None, 'max_retries': None, 'logprobs': None, 'top_logprobs': None, 'extra_headers': None, 'api_version': None, 'parallel_tool_calls': None, 'drop_params': None, 'allowed_openai_params': None, 'reasoning_effort': None, 'additional_drop_params': None, 'messages': [{'role': 'system', 'content': 'You are a helpful assistant.'}, {'role': 'user', 'content': 'Write a haiku about the ocean burning'}], 'thinking': None, 'web_search_options': None}
18:06:08 - LiteLLM:DEBUG: utils.py:3236 - 
LiteLLM: Non-Default params passed to completion() {}
18:06:08 - LiteLLM:DEBUG: utils.py:348 - Final returned optional params: {'extra_body': {}}
18:06:08 - LiteLLM:DEBUG: litellm_logging.py:475 - self.optional_params: {'extra_body': {}}
18:06:08 - LiteLLM:DEBUG: utils.py:4592 - checking potential_model_names in litellm.model_cost: {'split_model': 'gpt-4.1-nano', 'combined_model_name': 'openai/gpt-4.1-nano', 'stripped_model_name': 'gpt-4.1-nano', 'combined_stripped_model_name': 'openai/gpt-4.1-nano', 'custom_llm_provider': 'openai'}
18:06:08 - LiteLLM:DEBUG: litellm_logging.py:923 - 

POST Request Sent from LiteLLM:
curl -X POST \
https://llm \
-d '{'model': 'gpt-4.1-nano', 'messages': [{'role': 'system', 'content': 'You are a helpful assistant.'}, {'role': 'user', 'content': 'Write a haiku about the ocean burning'}], 'extra_body': {}}'

Relevant log output

Are you a ML Ops Team?

No

What LiteLLM version are you on ?

v1.74.8

Twitter / LinkedIn details

@emersongomesma