FastMCP调用工具时Context日志无法在控制台显示的问题

问题描述 #

按照FastMCP文档使用Context对象在工具中记录日志，工具调用正常，但服务器端仅能看到API访问日志，工具内通过ctx.debug()/ctx.info()等输出的日志完全不显示。使用版本：FastMCP 2.2.10。

工具代码：

from fastmcp import FastMCP, Context
from typing import Dict, Any, Optional

server = FastMCP("Demo")

@server.tool()
async def analyze_data(data: list[float], ctx: Context) -> dict:
    """Analyze numerical data with logging."""
    await ctx.debug("Starting analysis of numerical data")
    await ctx.info(f"Analyzing {len(data)} data points")

    try:
        result = sum(data) / len(data)
        await ctx.info(f"Analysis complete, average: {result}")
        return {"average": result, "count": len(data)}
    except ZeroDivisionError:
        await ctx.warning("Empty data list provided")
        return {"error": "Empty data list"}
    except Exception as e:
        await ctx.error(f"Analysis failed: {str(e)}")
        raise

客户端调用日志显示工具已成功执行：

{
'messages': [
     HumanMessage(content='Analyse this data 1.0, 2.01', additional_kwargs={}, response_metadata={}, id='45085e8f-2e39-4ef2-83a8-3958309bd92d'),
     
     AIMessage(content="I'll analyze the numerical data you've provided: 1.0 and 2.01. Let me do that for you.", additional_kwargs={'tool_calls': [{'id': 'toolu_bdrk_01URHoQ8iZ3BF5xsvB9osB6Z', 'function': {'arguments': '{"data":[1.0,2.01]}', 'name': 'analyze_data'}, 'type': 'function'}], 'refusal': None}, response_metadata={'token_usage': {'completion_tokens': 91, 'prompt_tokens': 2398, 'total_tokens': 2489, 'completion_tokens_details': None, 'prompt_tokens_details': None}, 'model_name': 'us.anthropic.claude-3-7-sonnet-20250219-v1:0', 'system_fingerprint': None, 'id': 'msg_bdrk_01G3NGwkfV5KEMD71DU9iTTX', 'service_tier': None, 'finish_reason': 'tool_calls', 'logprobs': None}, id='run--d75ffdcb-0dc1-479d-acb8-5f19fac793a8-0', tool_calls=[{'name': 'analyze_data', 'args': {'data': [1.0, 2.01]}, 'id': 'toolu_bdrk_01URHoQ8iZ3BF5xsvB9osB6Z', 'type': 'tool_call'}], usage_metadata={'input_tokens': 2398, 'output_tokens': 91, 'total_tokens': 2489, 'input_token_details': {}, 'output_token_details': {}}), 

     ToolMessage(content='{\n  "average": 1.505,\n  "count": 2\n}', name='analyze_data', id='13ce1458-e927-4985-8a8b-71580b0d5791', tool_call_id='toolu_bdrk_01URHoQ8iZ3BF5xsvB9osB6Z'), 

     AIMessage(content="Here's the analysis of the numerical data you provided:\n- Data points: 1.0, 2.01\n- Count: 2 data points\n- Average (mean): 1.505\n\nThis is a simple statistical summary of the two values you provided. If you'd like more detailed analysis, please let me know.", additional_kwargs={'refusal': None}, response_metadata={'token_usage': {'completion_tokens': 95, 'prompt_tokens': 2518, 'total_tokens': 2613, 'completion_tokens_details': None, 'prompt_tokens_details': None}, 'model_name': 'us.anthropic.claude-3-7-sonnet-20250219-v1:0', 'system_fingerprint': None, 'id': 'msg_bdrk_01R4j4Bycdz3SkeoG2rGySfL', 'service_tier': None, 'finish_reason': 'stop', 'logprobs': None}, id='run--ec20a5ac-5626-4938-9954-b9ef9dd29d73-0', usage_metadata={'input_tokens': 2518, 'output_tokens': 95, 'total_tokens': 2613, 'input_token_details': {}, 'output_token_details': {}})]
}

解决方法 #

1. 调整全局日志级别 #

FastMCP默认日志级别可能高于你使用的级别（比如默认是INFO，DEBUG日志会被过滤），需要显式设置全局日志级别：

import logging
from fastmcp import FastMCP, Context

# 全局设置日志级别为DEBUG，确保所有级别的日志都能输出
logging.basicConfig(level=logging.DEBUG)

server = FastMCP("Demo")

# 工具代码保持不变
@server.tool()
async def analyze_data(data: list[float], ctx: Context) -> dict:
    # ... 原有代码 ...

2. 检查ASGI服务器日志配置 #

如果用uvicorn等ASGI服务器启动FastMCP，需要确保服务器本身的日志级别也配置正确，避免日志被二次过滤。启动时加上--log-level debug参数：

uvicorn your_server_module:app --log-level debug

3. 排查Context日志转发问题 #

可以先在工具中直接使用标准logging模块打印日志，验证基础日志功能是否正常：

@server.tool()
async def analyze_data(data: list[float], ctx: Context) -> dict:
    logging.debug("Direct debug log from standard logging")
    await ctx.debug("Starting analysis of numerical data")
    # ... 原有代码 ...

如果标准日志能显示但Context日志不能，说明FastMCP的Context日志转发存在问题，建议：

• 升级FastMCP到最新版本，修复可能存在的版本bug
• 查看FastMCP源码确认Context日志的实现逻辑，是否需要额外配置

4. 确认FastMCP初始化时的日志参数 #

检查FastMCP初始化是否支持日志相关参数，部分版本可能需要显式传入日志配置：

server = FastMCP("Demo", log_level=logging.DEBUG)

内容的提问来源于stack exchange，提问作者mano_ksp