fix: honor custom timeout on every attempt in request_with_retry

[eeshsaxena]

Related Issues

No linked issue - found while reading the retry helper.

Proposed Changes:

request_with_retry and async_request_with_retry respect a custom timeout only on the first attempt; every retry silently falls back to the default of 10 seconds.

The cause is that timeout is popped from the shared kwargs dict inside the @retry-decorated inner run():

def run() -> httpx.Response:
    timeout = kwargs.pop("timeout", 10)   # runs on every attempt
    with httpx.Client() as client:
        res = client.request(**kwargs, timeout=timeout)

run() is re-invoked on each retry and kwargs is captured by closure, so the first attempt removes timeout from kwargs and subsequent attempts get the 10 default. A caller doing request_with_retry(url=..., timeout=30, attempts=3) gets timeout=30 on attempt 1 and timeout=10 on the retries.

The fix pops timeout once, before the retry loop, so it is captured for every attempt. Applied to both the sync and async functions.

How did you test it?

• Added regression tests to test/utils/test_requests_utils.py (sync and async): the first attempt fails with a retryable error, and the test asserts both the initial attempt and the retry are called with the caller's timeout, not the default. These fail on the current code ([5, 10]) and pass with the fix ([5, 5]).
• The existing test_..._custom_timeout only asserted the first attempt (assert_called_once_with), which is why this slipped through.
• Verified empirically by mocking httpx.Client.request to record the per-attempt timeout: current [5, 10, 10] for timeout=5, attempts=3, fixed [5, 5, 5]. Same for the async path.
• Full file passes (18 tests); ruff check and ruff format clean. Added a release note under releasenotes/notes/.

Notes for the reviewer

No source behavior changes beyond the timeout handling; the only functional change is that retries now use the caller's timeout. An AI assistant helped identify and implement this fix; I reviewed all changes and ran the tests locally.

Checklist

• I have read the contributors guidelines and the code of conduct.
• I have updated the related issue with new insights and changes.
• I have added unit tests and updated the docstrings.
• I've used one of the conventional commit types for my PR title.
• I have documented my code.
• I have added a release note file.
• I have run pre-commit hooks and fixed any issue.

[vercel]

@eeshsaxena is attempting to deploy a commit to the deepset Team on Vercel.

A member of the Team first needs to authorize it.

[CLAassistant]

All committers have signed the CLA.

[github-actions]

Coverage report

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
haystack/core/pipeline
pipeline.py
haystack/utils
requests_utils.py
Project Total

This report was generated by python-coverage-comment-action