This fork of the popular requests library introduces a new feature that allows you to specify the number of retry attempts via a retries parameter. This parameter is available in the following methods:
requests.get()requests.head()requests.post()requests.patch()requests.put()requests.options()requests.request()
pip install -e git+https://github.com/odinmckenzie/requests.git#egg=requestsYou can use the retries parameter with an integer value to indicate the number of retry attempts.
The following example will attempt to make a GET request up to 3 times before giving up.
import requests
from requests.exceptions import ConnectTimeout
try:
response = requests.get('http://10.1.1.1:8765/', timeout=0.25, retries=3)
except ConnectTimeout as e:
# for retries=3 it will send 4 requests
print("num_requests_sent: ", e.args[0].pool.num_requests)Expected output:
num_requests_sent: 4For more fine-grained control, you can pass in a Retry object from urllib3.util.retry. This allows you to specify different policies for different types of errors. In this example, the request will be retried up to 3 times for the specified HTTP status codes, with a backoff factor of 1 between each retry.
import requests
from urllib3.util.retry import Retry
from requests.exceptions import RetryError
# Create a Retry object
retry_strategy = Retry(
total=3, #number of retries
backoff_factor=1,
status_forcelist=[500, 502, 503, 504],
)
try:
response = requests.get('http://httpbin.org/status/503', retries=retry_strategy)
except RetryError as e:
# for total=3 it will send 4 requests
print("num_requests_sent: ", e.args[0].pool.num_requests)Expected output:
num_requests_sent: 4The retries parameter can also be used with the following methods of the Session object from requests.sessions:
get()head()post()patch()put()request()
The example will attempt to make a GET request up to 3 times before giving up.
import requests
from requests.exceptions import ConnectTimeout
from requests.sessions import Session
session = Session()
try:
response = session.request('GET', 'http://10.1.1.1:8765/', timeout=0.25, retries=3)
except ConnectTimeout as e:
# for retries=3 it will send 4 requests
print("num_requests_sent: ", e.args[0].pool.num_requests)Expected output:
num_requests_sent: 4The documentation for Requests suggested a solution for retrying failed connections. The solution shown below still works.
import requests
from requests import Session
from requests.adapters import HTTPAdapter
from requests.exceptions import RetryError
from urllib3.util.retry import Retry
s = Session()
retries = Retry(
total=3,
backoff_factor=0.1,
status_forcelist=[502, 503, 504],
allowed_methods={'GET'},
)
s.mount('https://', HTTPAdapter(max_retries=retries))
try:
response = s.get('https://httpbin.org/status/503')
except RetryError as e:
# for retries=3 it will send 4 requests
print("num_requests_sent: ", e.args[0].pool.num_requests)The methods post() and patch() only retry on a connection error and not for any specified HTTP status codes. This is because of implementation details in the urllib3 library.