httpx Drop `backend=...` argument. (Fix)

The Fix

Upgrade to version 0.13.0 or later.

Based on closed encode/httpx issue #782 · 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

@@ -95,18 +95,9 @@ await client.post(url, data=upload_bytes())
 HTTPX supports either `asyncio` or `trio` as an async environment.
 
-By default it will auto-detect which of those two to use as the backend
+It will auto-detect which of those two to use as the backend
 for socket operations and concurrency primitives.

repro.py

from httpx._backend.auto import AutoBackend
from httpx._dispatch.connection_pool import ConnectionPool

class CustomBackend(AutoBackend):
    async def open_tcp_stream(
        self,
        hostname: str,
        port: int,
        ssl_context: typing.Optional[ssl.SSLContext],
        timeout: Timeout,
    ) -> BaseSocketStream:
        value = await super().open_tcp_stream(hostname, port, ssl_context, timeout)
        # use value.stream_reader._transport.get_extra_info('ssl_object')
        return value

async def f(*args, **kwargs):
    backend = CustomBackend()
    dispatch = ConnectionPool(backend=backend, http2=True)
    async with AsyncClient(dispatch=dispatch) as client:
       ....

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 0.13.0 or later.\nWhen NOT to use: This fix is not suitable for users relying on custom backend implementations.\n\n

Why This Fix Works in Production

Trigger: This backend behaves as a proxy to a real backend, but record the ssl.SSLObject on the way.
Mechanism: The `backend` argument was unnecessary due to async autodetection
Why the fix works: Dropped the `backend` parameter from `AsyncClient.__init__`, simplifying the API as autodetection is now the default behavior. (first fixed release: 0.13.0).

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

The `backend` argument was unnecessary due to async autodetection
Production symptom (often without a traceback): This backend behaves as a proxy to a real backend, but record the ssl.SSLObject on the way.

Proof / Evidence

GitHub issue: #782
Fix PR: https://github.com/encode/httpx/pull/791
First fixed release: 0.13.0
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.85
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.55

Discussion

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

Jump to Sources Open on GitHub

“Note: I use a custom Backend to record the certificates of the HTTPS requests”

@dalf · 2020-01-23 · source

“@dalf - That's a really useful (surprising) bit of feedback, thanks! (And yes, you've got the right approach there, but make sure you're pinning your…”

@lovelydinosaur · 2020-01-24 · source

Failure Signature (Search String)

This backend behaves as a proxy to a real backend, but record the ssl.SSLObject on the way.
* https://github.com/dalf/searx-stats2/blob/master/searxstats/common/ssl_info.py#L85

Copy-friendly signature

signature.txt

Failure Signature
-----------------
This backend behaves as a proxy to a real backend, but record the ssl.SSLObject on the way.
* https://github.com/dalf/searx-stats2/blob/master/searxstats/common/ssl_info.py#L85

Error Message

Signature-only (no traceback captured)

error.txt

Error Message
-------------
This backend behaves as a proxy to a real backend, but record the ssl.SSLObject on the way.
* https://github.com/dalf/searx-stats2/blob/master/searxstats/common/ssl_info.py#L85

Minimal Reproduction

repro.py

from httpx._backend.auto import AutoBackend
from httpx._dispatch.connection_pool import ConnectionPool

class CustomBackend(AutoBackend):
    async def open_tcp_stream(
        self,
        hostname: str,
        port: int,
        ssl_context: typing.Optional[ssl.SSLContext],
        timeout: Timeout,
    ) -> BaseSocketStream:
        value = await super().open_tcp_stream(hostname, port, ssl_context, timeout)
        # use value.stream_reader._transport.get_extra_info('ssl_object')
        return value

async def f(*args, **kwargs):
    backend = CustomBackend()
    dispatch = ConnectionPool(backend=backend, http2=True)
    async with AsyncClient(dispatch=dispatch) as client:
       ....

What Broke

Users experienced confusion and potential misuse of the `backend` argument.

Why It Broke

The `backend` argument was unnecessary due to async autodetection

Fix Options (Details)

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

Upgrade to version 0.13.0 or later.

When NOT to use: This fix is not suitable for users relying on custom backend implementations.

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

Option D — Guard side-effects with OnceOnly Guardrail for side-effects

Mitigate duplicate external side-effects under retries/timeouts/agent loops by gating the operation before calling external systems.

Place OnceOnly between your code/agent and real side-effects (Stripe, emails, CRM, APIs).
Use a stable key per side-effect (e.g., customer_id + action + idempotency_key).
Fail-safe: configure fail-open vs fail-closed based on blast radius and spend risk.

Show example snippet (optional)

onceonly.py

from onceonly import OnceOnly
import os

once = OnceOnly(api_key=os.environ["ONCEONLY_API_KEY"], fail_open=True)

# Stable idempotency key per real side-effect.
# Use a request id / job id / webhook delivery id / Stripe event id, etc.
event_id = "evt_..."  # replace
key = f"stripe:webhook:{event_id}"

res = once.check_lock(key=key, ttl=3600)
if res.duplicate:
    return {"status": "already_processed"}

# Safe to execute the side-effect exactly once.
handle_event(event_id)

See OnceOnly SDK

When NOT to use: Do not use this to hide logic bugs or data corruption. Use it to block duplicate external side-effects and enforce tool permissions/spend caps.

Fix reference: https://github.com/encode/httpx/pull/791

First fixed release: 0.13.0

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

When NOT to Use This Fix

This fix is not suitable for users relying on custom backend implementations.
Do not use this to hide logic bugs or data corruption. Use it to block duplicate external side-effects and enforce tool permissions/spend caps.

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

Add a CI check that diffs key outputs after upgrades (OpenAPI schema snapshots, JSON payload shapes, CLI output).
Upgrade behind a canary and run integration tests against the canary before 100% rollout.
Add a TLS smoke test that performs a real handshake in CI (include CA bundle validation and hostname checks).
Alert on handshake failures by error string and endpoint to catch cert/CA changes quickly.

Version Compatibility Table

Version	Status
0.13.0	Fixed

Related Issues

No related fixes found.

Cluster: httpx:configuration-error All hubs All clusters

Related clusters: Data consistency Other Timeout

Sources

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