fetch-contributor-data: Use builtin backoff.

This correctly handles connection timeouts, not just non-200 status
codes.

tools/fetch-contributor-data

@@ -10,8 +10,6 @@ import os
 import sys
 import unicodedata
 from datetime import date
-from random import randrange
-from time import sleep
 from typing import Dict, List, Optional, Union
 
 sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
@@ -21,8 +19,8 @@ setup_path()
 os.environ["DJANGO_SETTINGS_MODULE"] = "zproject.settings"
 
 import django
-import requests
 from django.conf import settings
+from requests.packages.urllib3.util.retry import Retry
 from typing_extensions import TypedDict
 
 django.setup()
@@ -58,7 +56,6 @@ logger = logging.getLogger("zulip.fetch_contributors_json")
 
 def fetch_contributors(repo_name: str, max_retries: int) -> List[Contributor]:
     contributors: List[Contributor] = []
-    retry_attempts = 0
     page_index = 1
 
     api_link = f"https://api.github.com/repos/zulip/{repo_name}/contributors"
@@ -70,29 +67,33 @@ def fetch_contributors(repo_name: str, max_retries: int) -> List[Contributor]:
     if personal_access_token is not None:
         headers = {"Authorization": f"token {personal_access_token}"}
 
-    session = GithubSession()
+    Retry.BACKOFF_MAX = 64
+    retry = Retry(
+        total=max_retries,
+        backoff_factor=2.0,
+        status_forcelist=set(
+            [
+                403,  # Github does unauth rate-limiting via 403's
+                429,  # The formal rate-limiting response code
+                502,  # Bad gateway
+                503,  # Service unavailable
+            ]
+        ),
+    )
+    session = GithubSession(max_retries=retry)
     while True:
-        response: requests.Response = session.get(
+        response = session.get(
             api_link,
             params={**api_data, "page": f"{page_index}"},
             verify=certificates,
             headers=headers,
         )
-        if response.status_code == 200:
-            data = response.json()
-            if len(data) == 0:
-                return contributors
-            contributors.extend(data)
-            retry_attempts = 0
-            page_index += 1
-        else:
-            retry_attempts += 1
-            if retry_attempts > args.max_retries:
-                logger.warning("Failed retries fetching contributors data from GitHub.")
-                sys.exit(1)
-
-            sleep_time = randrange(0, min(64, 2 ** retry_attempts))
-            sleep(sleep_time)
+        response.raise_for_status()
+        data = response.json()
+        if len(data) == 0:
+            return contributors
+        contributors.extend(data)
+        page_index += 1
 
 
 def write_to_disk(json_data: ContributorsJSON, out_file: str) -> None:

zerver/lib/github.py

@@ -1,5 +1,6 @@
 import json
 import logging
+from typing import Any
 
 import requests
 
@@ -10,8 +11,8 @@ logger = logging.getLogger(__name__)
 
 
 class GithubSession(OutgoingSession):
-    def __init__(self) -> None:
-        super().__init__(role="github", timeout=5)
+    def __init__(self, **kwargs: Any) -> None:
+        super().__init__(role="github", timeout=5, **kwargs)
 
 
 def get_latest_github_release_version_for_repo(repo: str) -> str: