Infrastructure improvements:

• Consolidate to ModelPool (remove redundant BackendPool)
• Add ConnectionMode, WorkerType, Runtime enums for type safety
• Add pytest_collection_modifyitems hook to scan test markers and auto-detect required backends (grpc/http)
• Simplify backends.py to only handle cloud backends (openai, xai)
• Add process_utils.py with common process management functions

Test additions:

• Add MMLU evaluation tests in e2e_test/router/
• Add pyproject.toml for e2e_test with pytest-rerunfailures

CI updates:

• Add router-e2e-tests job to pr-test-rust.yml workflow
• Uses 4-gpu-a10 runner with retry support for flaky tests

Cleanup:

• Remove custom pytest_retry decorator (use pytest-rerunfailures)
• Delete unused fixtures (generate_test_certs.py, mock_worker.py)

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.
• Follow the SGLang code style guidance.

Review Process

1. Ping Merge Oncalls to start the PR flow. See the PR Merge Process.
2. Get approvals from CODEOWNERS and other reviewers.
3. Trigger CI tests with comments (/tag-run-ci-label, /rerun-failed-ci, /tag-and-rerun-ci) or contact authorized users to do so.
4. After green CI and required approvals, ask Merge Oncalls to merge.