Codecov Report

❌ Patch coverage is 4.25532% with 45 lines in your changes missing coverage. Please review.
✅ Project coverage is 79.30%. Comparing base (a6bff9a) to head (d08beea).
⚠️ Report is 4 commits behind head on main.

@@            Coverage Diff             @@
##             main    #2676      +/-   ##
==========================================
- Coverage   79.37%   79.30%   -0.07%     
==========================================
  Files         662      662              
  Lines       52526    52590      +64     
  Branches      734      734              
==========================================
+ Hits        41690    41707      +17     
- Misses      10756    10803      +47     
  Partials       80       80              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

What there a reason you didn't use Retry from urllib3?

import requests
from requests.adapters import HTTPAdapter
from urllib3.util import Retry

session = requests.Session()
retries = Retry(total=5, backoff_factor=1, status_forcelist=[429, 500, 502, 503])
session.mount('http://', HTTPAdapter(max_retries=retries))

response = session.get('http://example.com')

What there a reason you didn't use Retry from urllib3?

import requests
from requests.adapters import HTTPAdapter
from urllib3.util import Retry

session = requests.Session()
retries = Retry(total=5, backoff_factor=1, status_forcelist=[429, 500, 502, 503])
session.mount('http://', HTTPAdapter(max_retries=retries))

response = session.get('http://example.com')

Retrying regular 500 is super dangerous, because it could duplicate actions (don't know where in the code it crashed). 502, 503 is probably ok.

What about Ruby?

Based on claude's analysis, faraday-retry cannot directly replace the current custom retry logic. Here's why:

The custom retry logic does something faraday-retry cannot:

  # Current logic between retries (lines 224-227 in json_api_object.rb):
  disconnect()       # Closes and nils @http
  sleep(RETRY_DELAY)
  connect()          # Creates a brand new Faraday.new() instance

Why faraday-retry Won't Work

1. faraday-retry is middleware - It operates within an existing Faraday connection and retries requests on the same connection. It cannot tear down and rebuild the connection object itself.
2. Connection state - When you get Errno::ECONNRESET, Errno::EPIPE, or IOError, the underlying TCP connection is likely in a bad state. The current code handles this by completely destroying and recreating the Faraday instance. faraday-retry would just retry on the potentially broken connection.
3. Custom hooks not supported - faraday-retry has retry_block callbacks, but these are for logging/monitoring, not for executing disconnect()/connect() to reset connection state.

We could use faraday-retry if we changed the architecture to not persist the @http connection - i.e., create a new Faraday instance for each request. But that would:

• Remove connection reuse (performance impact)
• Require significant refactoring

The same thing applies for the Python code and urllib3:

1. urllib3 Retry operates at the connection pool level - It's configured via requests.adapters.HTTPAdapter and manages retries within the underlying connection pool. It cannot tear down and recreate the requests.Session() object.
2. No callback/hook mechanism - urllib3's Retry class is purely declarative configuration (total, connect, read, backoff_factor, etc.). There's no way to inject disconnect()/connect() calls between retries.
3. Different scope - urllib3 Retry handles low-level connection pool retries. The current code needs application-level session management - completely destroying and recreating the Session object to get a fresh state.
4. Exception handling - The code catches ChunkedEncodingError, RequestsConnectionError, ConnectionResetError, BrokenPipeError, OSError. While urllib3 Retry handles some connection errors, it wouldn't help with the session recreation requirement.