fix: Forward HTTP headers for generate requests

[mudit-eng]

What does the PR do?

Forwards matching HTTP headers as inference request parameters in the /generate and /generate_stream HTTP paths by calling the same ForwardHeaders helper used by /infer.

Adds a generate endpoint regression using an ensemble that routes to the Python backend and verifies a forwarded HTTP header is visible through request.parameters().

Checklist

• PR title reflects the change and is of format <commit_type>: <Title>
• Changes are described in the pull request.
• Related issues are referenced.
• Populated github labels field
• Added test plan and verified test passes where the local runtime supports it.
• Verified that the PR passes existing CI.
• Verified copyright is correct on all changed files.
• Added succinct git squash message before merging ref.
• All template sections are filled out.
• Optional: Additional screenshots for behavior/output changes with before/after.

Commit Type:

• build
• ci
• docs
• feat
• fix
• perf
• refactor
• revert
• style
• test

Related PRs:

None.

Where should the reviewer start?

src/http_server.cc, then qa/L0_http/generate_endpoint_test.py.

Test plan:

• git diff --check HEAD~1..HEAD

• python3 -m py_compile qa/L0_http/generate_endpoint_test.py qa/python_models/generate_models/mock_llm/1/model.py

• qa/L0_http/test.sh generate endpoint section (not run locally: /opt/tritonserver/bin/tritonserver is not present in this environment)

• CI Pipeline ID:
  N/A

Caveats:

Local end-to-end QA could not be run because this environment does not include the Triton server runtime binary.

Background

Linear issue TGH-116 reported that --http-header-forward-pattern worked for /infer but not /generate when routing from an ensemble to the Python backend.

Related Issues: (use one of the action keywords Closes / Fixes / Resolves / Relates to)

• Resolves Linear TGH-116

Linear Issue: TGH-116