[RFC]: Add E2E full test for Qwen3-Omni-30B-A3B-Instruct

[yenuo26]

Motivation.

Build E2E CI for vllm-omni to strengthen quality protection. This update ensures robust validation for both online (real-time inference) and offline (batch/development) scenarios.

Proposed Change.

We will add end-to-end test cases for the Qwen3-Omni-30B-A3B-Instruct model.
The test plan and test cases are visible in Section 1 and 2 of the design document.

Implementation Roadmap & To-Do

Phase 1: Test Single Modal + a few request for online

• Submit Two Text case as template [Test] Add full test for Qwen3-Omni-30B-A3B-Instruct #720
• Submit the remaining test cases in this phase for image and audio modal [Test] Add full test for Qwen3-Omni-30B-A3B-Instruct for image and audio single modal #827
• Submit the remaining test cases in this phase for video modal Add video processing tests for e2e-online #892
• include the corresponding labels. [CI] Add pytest markers to current tests and update the doc. #577
• Submit files related to pipeline construction [Test] Add BuildKite test-full script for full CI. #867
• Extract a common instance startup method to avoid launching it for each test case.

---

Phase 2: Test mix Modal + a few request for online

• Submit the test cases for this phase

---

Phase 3: Test mix Modal + A large number of requests for online

• Submit vllm online benchmark
• Submit the test cases for this phase

---

Phase 4: Test offline

• Submit the test cases for this phase

---

Phase 5: Test chunk prefill feature for online

• Submit the test cases for this phase

Feedback Period.

No response

CC List.

No response

Any Other Things.

No response

Before submitting a new issue...

• Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.