urllib3 中重试之间的静态时间

我想在重试之间使用静态时间，而不是 urllib3 Retry 类中的指数退避因子。

我尝试包装 Retry 类来实现静态重试。然而，虽然 LoggingRetry 正在使用适当的参数进行实例化，但在实际重试发生时将使用默认值。在下面的示例中，重试之间应该有两秒钟的时间。 backoff_factor 会持续存在，但我似乎无法让 delay_between_retries 坚持下去。

谢谢。

import time
import urllib3
from urllib3.util.retry import Retry
from urllib3.exceptions import MaxRetryError
from requests. exceptions import HTTPError
import traceback



# Define the retry strategy with logging
class LoggingRetry(Retry):
    def __init__(self, delay_between_retries=0, *args, **kwargs):
        print(f"Initializing LoggingRetry with delay_between_retries={delay_between_retries}")
        self.delay_between_retries = delay_between_retries
        super().__init__(*args, **kwargs)

    def sleep(self, response=None):
        backoff = self.get_backoff_time()
        total_sleep_time = self.delay_between_retries + backoff
        print(f"LoggingRetry sleep -  delay_between_retries: {self.delay_between_retries}, Backoff time: {backoff}")

        time.sleep(total_sleep_time)


class MyClient:
    def __init__(self, total_tries=3, delay_between_retries=0, backoff_factor=0):
        self.total_tries = total_tries
        self.delay_between_retries = delay_between_retries
        self.backoff_factor = backoff_factor

        # Define a retry strategy with a custom fixed_wait_time
        self.retry_strategy = LoggingRetry(
            total=self.total_tries,
            status_forcelist=[202],
            backoff_factor=self.backoff_factor,
            delay_between_retries=self.delay_between_retries,
            raise_on_status=False,
            allowed_methods=frozenset(['POST'])
        )
        
        # Initialize HTTP client with retry strategy
        self.http = urllib3.PoolManager(retries=self.retry_strategy, maxsize=4, block=True)

    
    def request_data(self, url):    
        try:
            response = self.http.request('GET', url,)
            return response
        except MaxRetryError as me:
            print(f"Max retries reached: {me}")
        except HTTPError as he:
            print(f"An HTTP error occurred: {he}")
            traceback.print_exc()
        except Exception as e:
            print(f"An error occurred: {e}")
            traceback.print_exc()

MyClient(delay_between_retries=2, backoff_factor=0).request_data(url='http://0.0.0.202/')

当然可以，我可以帮解决这个问题。对 urllib3 重试机制的理解是正确的，并且的 LoggingRetry 类中几乎实现了期望的行为。问题出在对 get_backoff_time() 方法的调用中。

在 urllib3 的 Retry 类中， sleep() 方法（已经正确地对其进行了覆盖）默认情况下使用 self.get_backoff_time() 来计算睡眠时间。此方法引入了指数后退行为。

为了实现 静态 重试时间间隔，需要调整 LoggingRetry 类的 sleep() 方法，使其不使用 get_backoff_time() 计算值。

以下是经过修改的 LoggingRetry 类：

class LoggingRetry(Retry):
    def __init__(self, delay_between_retries=0, *args, **kwargs):
        print(f"Initializing LoggingRetry with delay_between_retries={delay_between_retries}")
        self.delay_between_retries = delay_between_retries
        super().__init__(*args, **kwargs)

    def sleep(self, response=None):
        # 删除以下行以禁用指数后退计算：
        # backoff = self.get_backoff_time() 
        total_sleep_time = self.delay_between_retries
        print(f"LoggingRetry sleep -  delay_between_retries: {self.delay_between_retries}, Total sleep time: {total_sleep_time}")

        time.sleep(total_sleep_time)

通过此更改， sleep() 方法现在将始终使用在 delay_between_retries 参数中传递的值，从而在每次重试时提供一个静态的 2 秒延迟。

希望这对有所帮助！如果有任何其他问题，请告诉我。