httpx Async client does not close underlying stream thus

The Fix

Upgrade to version 0.15.3 or later.

Based on closed encode/httpx issue #1315 · 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

@@ -1505,7 +1505,7 @@ async def _send_single_request(
         async def on_close(response: Response) -> None:
             response.elapsed = datetime.timedelta(seconds=await timer.async_elapsed())
-            if hasattr(stream, "close"):
+            if hasattr(stream, "aclose"):
                 await stream.aclose()

repro.py

$ python3.8 -m asyncio
>> import asyncio
>> import httpx
>> c = httpx.AsyncClient()
>> await c.get('http://httpbin.org/')
<Response [200 OK]>
>> c._transport._connections
{(b'http', b'httpbin.org', 80): {<AsyncHTTPConnection http_version=HTTP/1.1 state=2>}}
>> await c.get('http://httpbin.org/')
<Response [200 OK]>
>> c._transport._connections
{(b'http', b'httpbin.org', 80): {<AsyncHTTPConnection http_version=HTTP/1.1 state=2>, <AsyncHTTPConnection http_version=HTTP/1.1 state=2>}}

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 0.15.3 or later.\nWhen NOT to use: Do not apply this fix if using a version of httpx that does not support async streams.\n\n

Why This Fix Works in Production

Trigger: Async client does not close underlying stream thus leaving connection active which leads to hitting pool limit
Mechanism: AsyncClient fails to close httpcore's stream after requests, causing connection pool limits to be reached
Why the fix works: Fixes the issue where the AsyncClient does not properly close the underlying stream, leading to connection pool limits being reached. (first fixed release: 0.15.3).

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 3.8 in real deployments (not just unit tests).
AsyncClient fails to close httpcore's stream after requests, causing connection pool limits to be reached
Production symptom (often without a traceback): Async client does not close underlying stream thus leaving connection active which leads to hitting pool limit

Proof / Evidence

GitHub issue: #1315
Fix PR: https://github.com/encode/httpx/pull/1316
First fixed release: 0.15.3
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.85
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.61

Discussion

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

Jump to Sources Open on GitHub

“### Checklist  - [x] The bug is reproducible against the latest release and/or master. - [x] There are no similar issues or pull requests to fix it yet. #”

Issue thread · issue description · source

Failure Signature (Search String)

Async client does not close underlying stream thus leaving connection active which leads to hitting pool limit
``AsyncClient`` doesn't close ``httpcore``'s stream after performing request, the bug shows itself in locking on connection pool's semaphore.

Copy-friendly signature

signature.txt

Failure Signature
-----------------
Async client does not close underlying stream thus leaving connection active which leads to hitting pool limit
``AsyncClient`` doesn't close ``httpcore``'s stream after performing request, the bug shows itself in locking on connection pool's semaphore.

Error Message

Signature-only (no traceback captured)

error.txt

Error Message
-------------
Async client does not close underlying stream thus leaving connection active which leads to hitting pool limit
``AsyncClient`` doesn't close ``httpcore``'s stream after performing request, the bug shows itself in locking on connection pool's semaphore.

Minimal Reproduction

repro.py

$ python3.8 -m asyncio
>> import asyncio
>> import httpx
>> c = httpx.AsyncClient()
>> await c.get('http://httpbin.org/')
<Response [200 OK]>
>> c._transport._connections
{(b'http', b'httpbin.org', 80): {<AsyncHTTPConnection http_version=HTTP/1.1 state=2>}}
>> await c.get('http://httpbin.org/')
<Response [200 OK]>
>> c._transport._connections
{(b'http', b'httpbin.org', 80): {<AsyncHTTPConnection http_version=HTTP/1.1 state=2>, <AsyncHTTPConnection http_version=HTTP/1.1 state=2>}}

Environment

Python: 3.8

What Broke

Connection pool limits are reached, leading to request failures in long-running processes.

Why It Broke

AsyncClient fails to close httpcore's stream after requests, causing connection pool limits to be reached

Fix Options (Details)

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

Upgrade to version 0.15.3 or later.

When NOT to use: Do not apply this fix if using a version of httpx that does not support async streams.

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

Fix reference: https://github.com/encode/httpx/pull/1316

First fixed release: 0.15.3

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

When NOT to Use This Fix

Do not apply this fix if using a version of httpx that does not support async streams.

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
0.15.3	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.

httpx Async client does not close underlying stream thus leaving connection active which leads to (Fix)