fix: respect user-configured timeout in OpenRouter requests

[yashwantbezawada]

What happened?

Fixes #16394

When setting a timeout for OpenRouter requests (e.g., timeout=10), the request does not timeout at 10 seconds. Instead, it continues for up to 300 seconds (aiohttp's default total timeout).

Root Cause

In aiohttp_transport.py, the ClientTimeout was created without setting the total field:

timeout=ClientTimeout(
    sock_connect=timeout.get("connect"),
    sock_read=timeout.get("read"),
    connect=timeout.get("pool"),
)

This meant that while socket-level timeouts were set, the overall request timeout was not, causing it to default to aiohttp's 300-second timeout.

Solution

Added the total parameter to ClientTimeout:

timeout=ClientTimeout(
    total=timeout.get("timeout"),  # ← Added this line
    sock_connect=timeout.get("connect"),
    sock_read=timeout.get("read"),
    connect=timeout.get("pool"),
)

Now when a user specifies timeout=10, the request will properly timeout after 10 seconds.

Impact

• Affected: OpenRouter users (and potentially other providers using base_llm_http_handler with aiohttp_transport)
• Severity: High - Users cannot reliably control request timeouts
• Backward compatibility: No breaking changes - only fixes existing broken behavior

Testing

The fix is straightforward and low-risk:

• Adds a single parameter that was missing
• Uses the same timeout dict structure already in use
• Does not change any other timeout behavior

[vercel]

@yashwantbezawada is attempting to deploy a commit to the CLERKIEAI Team on Vercel.

A member of the Team first needs to authorize it.

[CLAassistant]

All committers have signed the CLA.