时间戳

首页 > 解决方案 > 登录异步函数时如何使用 request_id？

python - 登录异步函数时如何使用 request_id？

问题描述

在异步函数中，每个 logger 语句都有自己的 request_id。

import logging
log = logging.getLogger('test_logger')

def sync_fun():
    log.info("test 1")
    log.info("test 2")
    log.info("test 3")

@after_response.enable
def async_fun():
    log.info("test 1")
    log.info("test 2")
    log.info("test 3")    

output of sync_fun:
[06/Nov/2019 10:42:00.234] [None] [130C6C47F1E24164AAC0440C719630] [INFO] Test 1
[06/Nov/2019 10:42:00.234] [None] [130C6C47F1E24164AAC0440C719630] [INFO] Test 2
[06/Nov/2019 10:42:00.234] [None] [130C6C47F1E24164AAC0440C719630] [INFO] Test 3

130C6C47F1E24164AAC0440C719630 是一个 request_id，它对所有记录器语句都很常见。

output of async_fun:
[06/Nov/2019 10:42:00.234] [None] [AB352B8F2DF9459ABDD2FBF51EB05F] [INFO] Test 1
[06/Nov/2019 10:42:00.234] [None] [V9E9B6DF5F9C442195EA7C1379FBFA] [INFO] Test 2
[06/Nov/2019 10:42:00.234] [None] [DCA311A92724443C9AD7E951288917] [INFO] Test 3

async_fun 是一个异步函数，所有记录器语句的请求 ID 都不同。

如何为异步函数中的每个记录器语句获取相同的 request_id。

我创建了日志过滤器。我在这里生成request_id。记录时使用相同的请求 ID。

from threading import local
_thread_locals = local()

class LoggingMiddleware(object):
    def process_request(self, request):
        if request.user.username:
            _thread_locals.user = request.user.username
            _thread_locals.user_email = request.user.email
            _thread_locals.user_id = request.user.id
        #assign request id to all requests (to track cron requests also)
        _thread_locals.request_id = str(uuid.uuid4().get_hex().upper()[0:30])


    def process_response(self, request, response):
        if hasattr(_thread_locals, 'user'):
            del _thread_locals.user
        if hasattr(_thread_locals, 'request_id'):
            del _thread_locals.request_id
        if hasattr(_thread_locals, 'user_email'):
            del _thread_locals.user_email
        if hasattr(_thread_locals, 'user_id'):
            del _thread_locals.user_id
        return response

    def process_exception(self, request, exception):
        logger.exception('unhandled error - ')

def get_current_user_details():
    user_details = {
        'username': getattr(_thread_locals, 'user', None),
        'email' : getattr(_thread_locals, 'user_email', None),
        'id' : getattr(_thread_locals, 'user_id', None)
    }
    return user_details

def get_current_user():
    return getattr(_thread_locals, 'user', None)

def get_current_request_id():
    return getattr(_thread_locals, 'request_id', None)

class RequestIDFilter(logging.Filter):
    def filter(self, record):
        current_user = get_current_user()
        current_request_id = get_current_request_id()
        record.user = current_user if current_user else None
        record.request_id = current_request_id if current_request_id else str(uuid.uuid4().get_hex().upper()[0:30])
        return True

标签： pythondjangoasynchronouslogging

解决方案

您可以使用单例模式，它将提供一个类的单个实例。

代码：

import logging

#Implement Singleton Pattern For Logger
class loggerSingleton(object):
   def __new__(myClass):
       if not hasattr(myClass, 'instance'):
           myClass.instance = super(loggerSingleton, myClass).__new__(myClass)
           myClass.instance.log = logging.getLogger('test_logger')
       return myClass.instance

#Let's try to create two instances      
singletonLog1 = loggerSingleton()
singletonLog2 = loggerSingleton()

#Let's check if we have single instances
print(singletonLog1.log is singletonLog2.log)

#print the logger ID's from two initializations
print(singletonLog1.log)
print(singletonLog2.log)

输出：

> True
> <logging.Logger object at 0x7f8995fcb6d8>
> <logging.Logger object at 0x7f8995fcb6d8>

大多数情况下，包括像你的情况一样的奇点，我更喜欢使用单例模式。

推荐阅读