如何优雅的在flask中记录log

时间 2019-11-10

标签如何优雅 flask 记录 log 栏目 Python 繁體版

原文原文链接

背景

记录日志，在任何项目中，都是很重要的。在Flask项目中，即有Flask提供的logger能够用来记录log，也能够经过直接使用Python的logging模块自定义logger来记录。那么这二者是什么关系，又该怎么使用呢？html

思路

Python的logging模块python

先看下对于logging模块的官方介绍git

Loggers have the following attributes and methods. Note that Loggers are never instantiated directly, but always through the module-level function logging.getLogger(name). Multiple calls to getLogger() with the same name will always return a reference to the same Logger object.
The name is potentially a period-separated hierarchical value, like foo.bar.baz (though it could also be just plain foo, for example). Loggers that are further down in the hierarchical list are children of loggers higher up in the list. For example, given a logger with a name of foo, loggers with names of foo.bar, foo.bar.baz, and foo.bam are all descendants of foo. The logger name hierarchy is analogous to the Python package hierarchy, and identical to it if you organise your loggers on a per-module basis using the recommended construction logging.getLogger(__name__). That’s because in a module, __name__ is the module’s name in the Python package namespace.github

https://docs.python.org/3/lib...shell

上面主要告诉咱们三点，flask
- 能够经过logging.getLogger(name)来获取一个logger，相同名字的logger，实际上是同一个logger。
- logger是经过name进行继承的，好比foo.bar就是foo 的子logger。就能够是实现咱们经过配置一个rootLogger，而后直接使用rootLogger.sublogger来记录一下内容，而不须要单独再配置一遍。
- 当使用logging.getLogger(__name__)时，__name__就是这个模块所在的python package的namespace。
flask提供的loggerapi

再看下flask中的logging模块：多线程

Flask uses standard Python logging. All Flask-related messages are logged under the 'flask' logger namespace. Flask.logger returns the logger named 'flask.app', and can be used to log messages for your application.
Depending on the situation, an extension may choose to log to app.logger or its own named logger. Consult each extension’s documentation for details.app

http://flask.pocoo.org/docs/1...socket

咱们能够知道flask的logger就是一个标准的Python logging，它的命名是flask。咱们既可使用app.logger，也能够本身定义一个logger。

那么如何使用app.logger呢？

有两种方式：
- 直接调用
```
logger = logging.getLogger('flask.app')
logger.info('flask.app')
```
- 使用Flask提供的接口
```
from flask import current_app
current_app.logger.info('logged by current_app from main')
```

这里推荐仍是使用第二种，current_app是一个单例，能够直接引用到app.logger。

经过修改app.logger的name，能够实现子logger的继承么？

答案是否认的。

修改app.logger的name：

# app/__init__.py
app.logger.name = 'app'

而后在子模块中定义一个app.module的logger来记录：

from flask import current_app
import logging

logger = logging.getLogger('app.module')

@module.route('/test', methods=['GET'])
def test():
    logger.info('logged by app.module')
    current_app.logger.info('logged by current_app.logger')

输出结果：

2019-02-01 10:56:01,877 - Thread-2 - app - INFO - logged by current_app.logger

只有current_app.logger的输出。

修改app.logger的name是否是无效呢？

咱们把子模块中的logger的name修改成flask.app.module：

from flask import current_app
import logging

logger = logging.getLogger('flask.app.module')

@module.route('/test', methods=['GET'])
def test():
    logger.info('logged by flask.app.module')
    current_app.logger.info('logged by current_app.logger')

输出结果：

2019-02-01 11:00:10,944 - Thread-2 - flask.app.module - INFO - logged by flask.app.module
2019-02-01 11:00:10,946 - Thread-2 - app - INFO - logged by current_app.logger

两个logger均输出了。

可见，经过修改app.logger.name能够在记录的时候显示为咱们设置的名称，但实际上这个logger仍是flask.app。

__name__的使用

在自定义logger的状况下，为了方便起见，咱们能够利用__name__这个参数。

前面说到：当使用logging.getLogger(__name__)时，__name__就是这个模块所在的python package的namespace。

通常Flask的工厂模式结构以下：

app
├── __init__.py
├── main
│   ├── __init__.py
│   ├── functions.py
│   └── views.py
└── module
    ├── __init__.py
    ├── functions.py
    └── views.py

那么咱们在先在app.__init__中定义rootLogger，而后再在app.module.functions.py中定义子Logger，均使用logging.getLogger(__name__):

# app.__init__.py 初始化rootlogger
rootLogger = logging.getLogger(__name__)
    rootLogger.setLevel(logging.DEBUG)
    socketHandler = logging.handlers.SocketHandler('localhost',logging.handlers.DEFAULT_TCP_LOGGING_PORT)
    rootLogger.addHandler(socketHandler)
    rootLogger.setLevel(logging.DEBUG)

# app.module.functions.py
import logging

logger = logging.getLogger(__name__)

def record_from_logging():
    logger.info('logged by logging from __name__')

输出：

2019-02-01 12:18:34,743 - MainThread - app - INFO - register root logger by __name__
2019-02-01 12:19:24,954 - Thread-4 - app.module.functions - INFO - logged by logging from __name__

能够发现输出的logger.name就是所在的文件目录，logger之间的继承关系与整个程序包保持一致。

总结

根据上面分析，那么怎么优雅的记录logger呢？

若是没有对模块进行分logger记录要求的话。能够直接使用在程序初始化的时候配置app.logger（能够自行设置logger.name）。在模块中经过import current_app来记录：

# app.__init__.py
def register_logging(app):
    app.logger.name = 'app'
    # logstash_handler
    stashHandler = logstash.LogstashHandler('app.config.get('ELK_HOST')', 'app.config.get('ELK_PORT')')
    app.logger.addHandler(stashHandler)

    # socket_handler
    socketHandler = logging.handlers.SocketHandler('localhost', logging.handlers.DEFAULT_TCP_LOGGING_PORT)
    app.logger.addHandler(socketHandler)
    
# app.module.function.py
from flask import current_app

@module.route('/test', methods=['GET'])
def test():
    current_app.logger.info('logging someting')
    return 'logged by current_app.logger'

输出效果：

2019-02-01 13:49:28,998 - Thread-2 - app - INFO - logged by current_app from main
2019-02-01 13:49:38,346 - Thread-3 - app - INFO - logged by current_app of functions

__注意__: 对于current_app.logger的引用不能经过以下方式，会有RuntimeError的报错。

from flask import current_app

logger = current_app.logger

## 异常
    raise RuntimeError(_app_ctx_err_msg)
RuntimeError: Working outside of application context.

This typically means that you attempted to use functionality that needed
to interface with the current application object in some way. To solve
this, set up an application context with app.app_context().  See the
documentation for more information.

若是但愿按本身的实际需求，对模块进行分logger记录要求的话。那么建议本身设置logger。

# app.__init__.py
def register_logging():
    # set own root logger
    rootLogger = logging.getLogger(__name__)
    rootLogger.setLevel(logging.DEBUG)
    # socketHandler
    socketHandler = logging.handlers.SocketHandler('localhost',logging.handlers.DEFAULT_TCP_LOGGING_PORT)
    rootLogger.addHandler(socketHandler)
    # logstash_handler
    stashHandler = logstash.LogstashHandler('app.config.get('ELK_HOST')', 'app.config.get('ELK_PORT')')
    rootLogger.addHandler(stashHandler)
    rootLogger.setLevel(logging.DEBUG)

# app.module.function.py
import logging

logger = logging.getLogger(__name__)

@module.route('/test', methods=['GET'])
def test():
    logger.info('logging someting')
    return 'logged by logging module'

输出效果：

2019-02-01 13:49:49,297 - Thread-5 - app.module.views - INFO - logged by flask.app.module
2019-02-01 13:50:01,013 - Thread-7 - app.module.functions - INFO - logged by logging module of functions

完整代码可参考：https://github.com/keejo125/flask_logging_demo

注意

关于python中logging的配置可参考官网：

https://docs.python.org/3/lib...

在配置handler时，常常会但愿日志能够按时间分割(TimedRotatingFileHandler)或者按大小分割(RotatingFileHandler).

可是在flask项目中，尤为开启多线程以后，在分割日志(doRollover())时会有文件读写的异常:

WindowsError: [Error 32]

建议使用SocketHandler，将日志发送给单独的LogServer来进行二次处理。

简易的接收socketlog的LogServer可参考：https://github.com/keejo125/f...

或者如今流行的stashHandler，将日志发送给ELK来进行二次处理。