retryhttp 0.1.0

Retry potentially-transient HTTP errors.

retryhttp

Retry potentially-transient HTTP errors.

Note: This project is in beta status. The API may change significantly.

Overview

Under the hood, retryhttp extends tenacity with custom retry and wait strategies, as well as a decorator that wraps tenacity.retry() with sensible defaults.

Supports both requests and httpx.

Quickstart

Install from git:

pip install git+https://github.com/austind/retryhttp.git@main

import httpx
from retryhttp import retry_http_errors

# See Default Behavior section below for what happens here
@retry_http_errors()
def get_example():
    response = httpx.get("https://example.com/")
    response.raise_for_status()

Default Behavior

By default, retryhttp.retry_http_errors() retries the following conditions:

• If the response status code is:
  • 429 Too Many Requests
  • 500 Internal Server Error
  • 502 Bad Gateway
  • 503 Service Unavailable
  • 504 Gateway Timeout
• If one of the following other exceptions is raised:
  • Network errors
    • httpx.ConnectError
    • httpx.ReadError
    • httpx.WriteError
    • requests.ConnectionError
  • Network timeouts
    • httpx.TimeoutError
    • requests.Timeout

Based on which error is raised, a different wait strategy is used:

• 429 Too Many Requests: Respect the Retry-After header, if present. If not, fall back to tenacity.wait_exponential()
• 5xx Server Errors: tenacity.wait_exponential_jitter()
• Network errors: tenacity.wait_exponential()
• Network timeouts: tenacity.wait_exponential_jitter()