Update E2E tests to include multiport case

[kfswain]

With the addition of multiport routing in: #1663

We have need to include additional e2e test coverage to ensure the EPP handles multiport correctly.

E2E test dir: https://github.com/kubernetes-sigs/gateway-api-inference-extension/tree/main/test/e2e/epp

This issue would be considered complete when:

• e2e test is updated to support multiport inferencePool
• tests cases added

[AvineshTripathi]

Hey @kfswain 👋, I’d like to work on this issue. Since this project is new to me, I’ve explored the relevant parts of the codebase and would appreciate a quick review before moving forward.

From my understanding:

I’ll need to update config/manifests/vllm/sim-deployment.yaml to include the argument --data-parallel-size=3.

Based on the llm-d-inference-sim README, I’ll also update test/testdata/inferencepool-e2e.yaml to add 8001 and 8002 under targetPorts.

As a final step, I plan to verify that these target ports are exposed on the Pod and functional by making /completion requests. This will likely happen in test/e2e/epp/e2e_test.go

The only part I’m unsure about is the best way to check whether each port is active and responding as expected. Could you please confirm if my approach sounds correct and share any suggestions for verifying multi-port availability?

Thanks