Challenge 2: Add Retry Logic — Possible Solution
====================================================================

# catalog/tasks.py

from celery import shared_task

@shared_task(
    bind=True,
    autoretry_for=(ConnectionError,),
    retry_backoff=True,
    max_retries=3,
)
def generate_book_report(self, book_id):
    from .models import Book
    book = Book.objects.get(pk=book_id)
    # Simulated external call that might raise ConnectionError,
    # e.g. calling a third-party pricing API to enrich the report:
    # external_data = fetch_market_price(book.title)
    print(f"Report: '{book.title}' — ${book.price}, in_stock={book.in_stock}")

WHY THIS WORKS
--------------
- bind=True gives the task access to `self` (the task instance itself),
  which is required for Celery's retry machinery to correctly track
  attempt counts, backoff timing, and task metadata across retries.
- autoretry_for=(ConnectionError,) tells Celery to automatically retry
  this specific task, WITHOUT any manual try/except or self.retry() call
  needed in the task body, whenever it raises a ConnectionError — other
  exception types are NOT automatically retried, so a genuine bug (like a
  Book.DoesNotExist from a bad book_id) still fails immediately rather
  than being retried pointlessly.
- retry_backoff=True spaces out retry attempts with increasing delays
  (roughly 1s, 2s, 4s by default) rather than retrying immediately three
  times in a row — useful when the failure is a temporary network blip
  that might need a few seconds to resolve, rather than hammering a
  struggling external service immediately again.
- max_retries=3 caps the total number of retry attempts — after three
  failed retries, Celery gives up and marks the task as permanently
  failed, rather than retrying forever on a genuinely broken dependency.