uvicorn Uvicorn fails to start in 0.10.1+ (Can't pickle

The Fix

Upgrade to version 0.11.4 or later.

Based on closed Kludex/uvicorn issue #478 · 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

@@ -294,9 +294,11 @@ def run(app, **kwargs):
     server = Server(config=config)
 
-    if config.reload and not isinstance(app, str):
+    if (config.reload or config.workers > 1) and not isinstance(app, str):
         logger = logging.getLogger("uvicorn.error")

repro.py

from logging.handlers import DatagramHandler

class Handler(DatagramHandler):
    def makePickle(self, record):
        # The "uvicorn.access" logger would attach the scope
        # with unpickleable object.
        record.scope = None
        return super().makePickle(record)

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\nUpgrade to version 0.11.4 or later.\nWhen NOT to use: This fix is not applicable if the application cannot be imported as a string.\n\n

Why This Fix Works in Production

Trigger: AttributeError: Can't pickle local object 'request_response.<locals>.app'
Mechanism: The application instance was being passed directly, causing pickling issues in multiprocessing
Why the fix works: Forces the application to be passed as an import string when using multiple workers, resolving the pickling issue. (first fixed release: 0.11.4).

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

Triggered by an upgrade/regression window: 0.10.1 breaks; 0.11.4 is the first fixed release.
Shows up under Python 3.6 in real deployments (not just unit tests).
The application instance was being passed directly, causing pickling issues in multiprocessing
Surfaces as: AttributeError: Can't pickle local object 'request_response.<locals>.app'

Proof / Evidence

GitHub issue: #478
Fix PR: https://github.com/kludex/uvicorn/pull/481
First fixed release: 0.11.4
Affected versions: 0.10.1
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.85
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.66

Discussion

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

Jump to Sources Open on GitHub

“Thanks for the prompt and detailed response! Changing the app instance parameter to the string style worked fine, and the comments about debug and reload…”

@kchernoff · 2019-11-12 · source

“Also while we're here, you may as well change debug=True to reload=True. https://github.com/encode/uvicorn/issues/480”

@lovelydinosaur · 2019-11-09 · source

“For those hitting the pickling error while using a custom handler, it's due to the app object in the record.scope. Strip it off solves the…”

@NivekNey · 2020-11-17 · source

“Thanks for raising this! Rather than importing the project, you'll need to use the string format for the app argument instead..”

@lovelydinosaur · 2019-11-09 · source

Failure Signature (Search String)

AttributeError: Can't pickle local object 'request_response.<locals>.app'

Error Message

Stack trace

error.txt

Error Message
-------------
AttributeError: Can't pickle local object 'request_response.<locals>.app'

Minimal Reproduction

repro.py

from logging.handlers import DatagramHandler

class Handler(DatagramHandler):
    def makePickle(self, record):
        # The "uvicorn.access" logger would attach the scope
        # with unpickleable object.
        record.scope = None
        return super().makePickle(record)

Environment

Python: 3.6

What Broke

Uvicorn fails to start, resulting in application downtime.

Why It Broke

The application instance was being passed directly, causing pickling issues in multiprocessing

Fix Options (Details)

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

Upgrade to version 0.11.4 or later.

When NOT to use: This fix is not applicable if the application cannot be imported as a string.

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

Fix reference: https://github.com/kludex/uvicorn/pull/481

First fixed release: 0.11.4

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

When NOT to Use This Fix

This fix is not applicable if the application cannot be imported as a string.

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
0.10.1	Broken
0.11.4	Fixed

Related Issues

No related fixes found.

Cluster: uvicorn:configuration-error All hubs All clusters

Related clusters: Data consistency Race condition SSL handshake

Sources

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

uvicorn Uvicorn fails to start in 0.10.1+ (Can't pickle local object) (Fix)