Python Logging Best Practice
v0.2.11
Chromatrace是一个专为高级日志记录功能而设计的Python软件包，包括跟踪和请求ID管理以及过程ID。它提供了灵活的日志记录配置，并支持彩色日志记录，以提高可见度。
我认为记录是任何应用程序的重要组成部分，并且至关重要的是拥有良好的结构化记录系统。 Chromatrace旨在提供一个简单易于使用的记录系统，可以集成到任何Python应用程序中。简而言之，色谱是在Python中登录的最佳实践。
您可以通过PIP安装色谱：
pip install chromatrace要在您的应用程序中使用色谱，您可以导入必要的组件：
from chromatrace import LoggingSettings , LoggingConfig , tracer配置您的记录设置：
logging_config = LoggingConfig (
settings = LoggingSettings (
application_level = "Development" ,
enable_tracing = True ,
ignore_nan_trace = False ,
enable_file_logging = True ,
)
)
logger = logging_config . get_logger ( __name__ )使用tracer装饰器跟踪您的功能：
@ tracer
async def my_async_function ():
logger . debug ( "Check something" )
logger . info ( "Doing something" )
logger . warning ( "Doing something" )
logger . error ( "Something went wrong" ) from lagom import Container
container = Container ()
from chromatrace import LoggingConfig , LoggingSettings
container [ LoggingSettings ] = LoggingSettings (
application_level = "Development" ,
enable_tracing = True ,
ignore_nan_trace = False ,
enable_file_logging = True ,
)
container [ LoggingConfig ] = LoggingConfig ( container [ LoggingSettings ])然后，将LoggingConfig添加到您的服务中：
import logging
from chromatrace import LoggingConfig
class SomeService :
def __init__ ( self , logging_config : LoggingConfig ):
self . logger = logging_config . get_logger ( self . __class__ . __name__ )
self . logger . setLevel ( logging . ERROR )
async def do_something ( self ):
self . logger . debug ( "Check something in second service" )
self . logger . info ( "Doing something in second service" )
self . logger . error ( "Something went wrong in second service" )结果：
[Development]-(2024-11-21 23:43:26)-[INFO]-[APIService]-FILENAME:api_app.py-FUNC:do_something-THREAD:MainThread-LINE:27 ::
Doing something in API service
[Development]-(2024-11-21 23:43:26)-[ERROR]-[APIService]-FILENAME:api_app.py-FUNC:do_something-THREAD:MainThread-LINE:28 ::
Something went wrong in API service
[T-dc1be4de]-[Development]-(2024-11-21 23:43:26)-[INFO]-[Main]-FILENAME:main.py-FUNC:main-THREAD:MainThread-LINE:21 ::
Starting main
[T-dc1be4de]-[Development]-(2024-11-21 23:43:26)-[ERROR]-[ExampleService]-FILENAME:example_service.py-FUNC:do_something-THREAD:MainThread-LINE:26 ::
Something went wrong
[T-dc1be4de]-[Development]-(2024-11-21 23:43:26)-[INFO]-[InnerService]-FILENAME:example_service.py-FUNC:do_something-THREAD:MainThread-LINE:13 ::
Doing something in second service
[T-dc1be4de]-[Development]-(2024-11-21 23:43:26)-[ERROR]-[InnerService]-FILENAME:example_service.py-FUNC:do_something-THREAD:MainThread-LINE:14 ::
Something went wrong in second service
[T-dc1be4de]-[Development]-(2024-11-21 23:43:26)-[DEBUG]-[AnotherSample]-FILENAME:sample.py-FUNC:do_something-THREAD:MainThread-LINE:12 ::
Check something
[T-dc1be4de]-[Development]-(2024-11-21 23:43:26)-[INFO]-[AnotherSample]-FILENAME:sample.py-FUNC:do_something-THREAD:MainThread-LINE:13 ::
Doing something
[T-dc1be4de]-[Development]-(2024-11-21 23:43:26)-[WARNING]-[AnotherSample]-FILENAME:sample.py-FUNC:do_something-THREAD:MainThread-LINE:14 ::
Doing something
[T-dc1be4de]-[Development]-(2024-11-21 23:43:26)-[ERROR]-[AnotherSample]-FILENAME:sample.py-FUNC:do_something-THREAD:MainThread-LINE:15 ::
Something went wrong
两个第一个日志是从跟踪中的，并且未添加跟踪ID中的日志消息。其余的日志在跟踪中，并且在日志消息中添加了跟踪ID -T T-dc1be4de 。
注意：重要的是每个类或服务都可以具有自己的日志级别。当您想拥有不同服务的日志级别时，这很有用。
from chromatrace import RequestIdMiddleware
app = FastAPI ()
app . add_middleware ( RequestIdMiddleware )结果：
[R-ffe0a9a2]-[Development]-(2024-11-22 00:13:53)-[INFO]-[APIService]-FILENAME:api_app.py-FUNC:read_root-THREAD:MainThread-LINE:38 ::
Hello World
[R-ffe0a9a2]-[Development]-(2024-11-22 00:13:53)-[ERROR]-[ExampleService]-FILENAME:example_service.py-FUNC:do_something-THREAD:MainThread-LINE:26 ::
Something went wrong
[R-ffe0a9a2]-[Development]-(2024-11-22 00:13:53)-[INFO]-[InnerService]-FILENAME:example_service.py-FUNC:do_something-THREAD:MainThread-LINE:13 ::
Doing something in second service
[R-ffe0a9a2]-[Development]-(2024-11-22 00:13:53)-[ERROR]-[InnerService]-FILENAME:example_service.py-FUNC:do_something-THREAD:MainThread-LINE:14 ::
Something went wrong in second service
如您所见，请求ID -R R-ffe0a9a2自动从处理请求的线程中添加到日志消息中。
from chromatrace import SocketRequestIdMiddleware
socket_application = SocketRequestIdMiddleware ( socket_application )完整示例可以在socket_app.py文件中找到。我建议您在做出任何决定之前先检查一下。客户端代码可以在socket_client.py文件中找到。
结果：
[S-4e2b7c5e]-[Development]-(2024-12-06 01:15:18)-[INFO]-[Socket]-FILENAME:socket_app.py-FUNC:connect-THREAD:MainThread-LINE:86 ::
Socket connected on main namespace. SID: wB-srwnv9Xa2w_8bAAAB
[S-4e2b7c5e]-[Development]-(2024-12-06 01:15:20)-[INFO]-[Socket]-FILENAME:socket_app.py-FUNC:message-THREAD:MainThread-LINE:90 ::
Received message on main namespace. SID: wB-srwnv9Xa2w_8bAAAB, Message: Hello from the client
[S-aaf46528]-[Development]-(2024-12-06 01:15:47)-[INFO]-[Socket]-FILENAME:socket_app.py-FUNC:connect-THREAD:MainThread-LINE:86 ::
Socket connected on main namespace. SID: FI3E_S_A-KsTi4RLAAAD
[S-aaf46528]-[Development]-(2024-12-06 01:15:49)-[INFO]-[Socket]-FILENAME:socket_app.py-FUNC:message-THREAD:MainThread-LINE:90 ::
Received message on main namespace. SID: FI3E_S_A-KsTi4RLAAAD, Message: Hello from the client
是的，插座日志也在跟踪中。跟踪ID- S-4e2b7c5e和S-aaf46528已添加到日志消息中。为了获得更好的体验，将前缀S添加到跟踪ID中，以将其与请求ID区分开。
from chromatrace . uvicorn import GetLoggingConfig , UvicornLoggingSettings
rest_application = FastAPI ()
uvicorn . run (
rest_application ,
host = "0.0.0.0" ,
port = 8000 ,
log_level = "debug" ,
log_config = GetLoggingConfig (
UvicornLoggingSettings (
enable_file_logging = True ,
show_process_id = True ,
)
),
)结果：
(2024-12-12 20:54:54)-[PID:3710345]-[INFO]: Started server process [3710345]
(2024-12-12 20:54:54)-[PID:3710345]-[INFO]: Waiting for application startup.
(2024-12-12 20:54:54)-[PID:3710345]-[INFO]: Application startup complete.
(2024-12-12 20:54:54)-[PID:3710345]-[INFO]: Uvicorn running on http://0.0.0.0:8000 (Press CTRL+C to quit)
(2024-12-12 20:54:57)-[PID:3710345]-[INFO]: ADDRESS:(127.0.0.1:46166) - REQUEST:"GET /consume HTTP/1.1" - STATUS:200 OK
(2024-12-12 20:55:45)-[PID:3710345]-[INFO]: ADDRESS:(127.0.0.1:54018) - REQUEST:"GET /consume HTTP/1.1" - STATUS:200 OK
(2024-12-12 20:56:51)-[PID:3710345]-[INFO]: ADDRESS:(127.0.0.1:58240) - REQUEST:"GET / HTTP/1.1" - STATUS:200 OK
(2024-12-12 20:56:51)-[PID:3710345]-[INFO]: ADDRESS:(127.0.0.1:58254) - REQUEST:"GET / HTTP/1.1" - STATUS:200 OK
(2024-12-12 20:56:52)-[PID:3710345]-[INFO]: ADDRESS:(127.0.0.1:58260) - REQUEST:"GET / HTTP/1.1" - STATUS:200 OK
(2024-12-12 20:56:52)-[PID:3710345]-[INFO]: ADDRESS:(127.0.0.1:58270) - REQUEST:"GET / HTTP/1.1" - STATUS:200 OK
(2024-12-12 21:16:45)-[PID:3710345]-[INFO]: Shutting down
(2024-12-12 21:16:45)-[PID:3710345]-[INFO]: Waiting for application shutdown.
(2024-12-12 21:16:45)-[PID:3710345]-[INFO]: Application shutdown complete.
(2024-12-12 21:16:45)-[PID:3710345]-[INFO]: Finished server process [3710345]
日志在过程ID中 - PID:3710345已添加到日志消息中。日志消息也有色，以更好地可见性。如果enable_file_logging设置为True ，则日志消息也将写入文件。有关更多信息，请检查config.py文件， UvicornLoggingSettings类。
你不信任我，对吗？我明白。您想在行动中看到它，对吗？我给你覆盖了。 :)
您可以在示例目录中找到如何使用色谱的示例。使用以下命令运行示例：
python main.py然后，运行：
curl 0.0.0.0:8000现在，检查终端中的日志。
此外，套接字服务器将启动并等待客户端在http://localhost:8001上连接。对于插座客户端，在另一个终端运行以下命令：
python adaptors/socket_client.py现在，检查两个终端中的日志。
该项目是根据GNU Affero通用公共许可证获得许可的 - 有关详细信息，请参见许可证文件。
