Celery autoretry_for and retry_kwargs for class-based tasks

The Fix

pip install celery==4.4.0

Based on closed celery/celery issue #4684 · PR/commit linked

Production note: Most teams hit this during upgrades or environment changes. Roll out with a canary and smoke critical endpoints (health, OpenAPI/docs) before 100%.

Jump to Verify Open PR/Commit

@@ -460,11 +460,24 @@ def _task_from_fun(self, fun, name=None, base=None, bind=False, **options):
             task.bind(self)  # connects task to this app
 
-            autoretry_for = tuple(options.get('autoretry_for', ()))
-            retry_kwargs = options.get('retry_kwargs', {})
-            retry_backoff = int(options.get('retry_backoff', False))

repro.py

from project.celery import app


class test_task(celery.Task):
    autoretry_for = (Exception,)
    retry_kwargs = {
        'max_retries': 5,
        'countdown': 10,
    }

    def run(self, *args, **kwargs):
        raise Exception('test')


test_task = app.register_task(test_task())

test_task.delay()

verify

Re-run the minimal reproduction on your broken version, then apply the fix and re-run.

fix.md

Option A — Upgrade to fixed release\npip install celery==4.4.0\nWhen NOT to use: This fix should not be used if class-based tasks are not intended to have retry functionality.\n\nOption C — Workaround\na class similar to this could help you:\nWhen NOT to use: This fix should not be used if class-based tasks are not intended to have retry functionality.\n\n

Why This Fix Works in Production

Trigger: autoretry_for = (Exception,)
Mechanism: The autoretry_for attribute was not applied to class-based tasks due to implementation oversight
Why the fix works: Allows class-based tasks to define retry attributes such as autoretry_for and retry_kwargs, addressing issue #4684. (first fixed release: 4.4.0).

Production impact:

If left unfixed, the same config can fail only in production (env differences), causing startup failures or partial feature outages.

Why This Breaks in Prod

Shows up under Python 3.7.4 in real deployments (not just unit tests).
The autoretry_for attribute was not applied to class-based tasks due to implementation oversight
Production symptom (often without a traceback): autoretry_for = (Exception,)

Proof / Evidence

GitHub issue: #4684
Fix PR: https://github.com/celery/celery/pull/5869
First fixed release: 4.4.0
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.85
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.63

Discussion

High-signal excerpts from the issue thread (symptoms, repros, edge-cases).

Jump to Sources Open on GitHub

“I think this would be a very neat feature (unless we want to actively discourage the use of class based tasks altogether) - if one…”

@pawelad · 2018-12-10 · source

“Exactly the same issue for me. It looks like the autoretry_for attribute is only touched in _task_from_fun, which isn't used for a class based task.…”

@amir-hadi · 2018-05-05 · source

“@CompadreP maybe as a little workaround a class similar to this could help you:”

@amir-hadi · 2018-05-05 · source

“@amir-hadi Thanks for the idea. With small modifications I've got working retries with exponential backoff. Celery 4.3.0 Python 3.7.4 Further improvements: * Make it a…”

@icamys · 2019-10-30 · source

Failure Signature (Search String)

autoretry_for = (Exception,)
raise Exception('test')

Copy-friendly signature

signature.txt

Failure Signature
-----------------
autoretry_for = (Exception,)
raise Exception('test')

Error Message

Signature-only (no traceback captured)

error.txt

Error Message
-------------
autoretry_for = (Exception,)
raise Exception('test')

Minimal Reproduction

repro.py

from project.celery import app


class test_task(celery.Task):
    autoretry_for = (Exception,)
    retry_kwargs = {
        'max_retries': 5,
        'countdown': 10,
    }

    def run(self, *args, **kwargs):
        raise Exception('test')


test_task = app.register_task(test_task())

test_task.delay()

Environment

Python: 3.7.4

What Broke

Class-based tasks did not retry on failure, leading to task execution failures.

Why It Broke

The autoretry_for attribute was not applied to class-based tasks due to implementation oversight

Fix Options (Details)

Option A — Upgrade to fixed release Safe default (recommended)

pip install celery==4.4.0

When NOT to use: This fix should not be used if class-based tasks are not intended to have retry functionality.

Use when you can deploy the upstream fix. It is usually lower-risk than long-lived workarounds.

Option C — Workaround Temporary workaround

a class similar to this could help you:

When NOT to use: This fix should not be used if class-based tasks are not intended to have retry functionality.

Use only if you cannot change versions today. Treat this as a stopgap and remove once upgraded.

Fix reference: https://github.com/celery/celery/pull/5869

First fixed release: 4.4.0

Last verified: 2026-02-09. Validate in your environment.

When NOT to Use This Fix

This fix should not be used if class-based tasks are not intended to have retry functionality.

Verify Fix

verify

Re-run the minimal reproduction on your broken version, then apply the fix and re-run.

Did This Fix Work in Your Case?

Quick signal helps us prioritize which fixes to verify and improve.

Prevention

Capture the exact failing error string in logs and tests so you can reproduce via a minimal script.
Pin production dependencies and upgrade only with a reproducible test that hits the failing path.

Version Compatibility Table

Version	Status
4.4.0	Fixed

Related Issues

No related fixes found.

Cluster: celery:configuration-error Celery hub Celery best practices All hubs All clusters

Related clusters: Data consistency Duplicates Duplicate processing

Sources

We don’t republish the full GitHub discussion text. Use the links above for context.

Celery autoretry_for and retry_kwargs for class-based tasks (Fix)