Python Mid-level (2–5 yrs) Coding Frequent Python decorator retry exponential backoff functools

Implement a retry decorator with configurable attempts and exponential backoff

py-mid-004

Your answer

Answer as you would in a real interview — explain your thinking, not just the conclusion.

Model answer

A retry decorator wraps a function and re-calls it on failure up to max_attempts times. Use functools.wraps to preserve the original function's metadata. Exponential backoff formula: sleep(base * 2^attempt). Add jitter (random fraction) to prevent thundering herd when many coroutines retry simultaneously. The decorator should accept configuration parameters, which requires a decorator factory (a function that returns a decorator). For async code, use asyncio.sleep instead of time.sleep.

Code example

import functools
import time
import random
from typing import Callable, TypeVar, Any

F = TypeVar("F", bound=Callable[..., Any])

def retry(
    max_attempts: int = 3,
    base_delay: float = 0.5,
    exceptions: tuple[type[Exception], ...] = (Exception,),
) -> Callable[[F], F]:
    """Decorator factory: retry on exception with exponential backoff + jitter."""
    def decorator(func: F) -> F:
        @functools.wraps(func)
        def wrapper(*args, **kwargs):
            for attempt in range(max_attempts):
                try:
                    return func(*args, **kwargs)
                except exceptions as exc:
                    if attempt == max_attempts - 1:
                        raise  # final attempt — re-raise
                    delay = base_delay * (2 ** attempt) + random.uniform(0, 0.1)
                    print(f"Attempt {attempt + 1} failed: {exc}. Retrying in {delay:.2f}s")
                    time.sleep(delay)
        return wrapper  # type: ignore[return-value]
    return decorator

# Usage
import requests

@retry(max_attempts=4, base_delay=1.0, exceptions=(requests.RequestException,))
def fetch(url: str) -> dict:
    response = requests.get(url, timeout=5)
    response.raise_for_status()
    return response.json()

Follow-up

How would you extend this to only retry on specific exception types, and how would you make it compatible with both sync and async functions?