Points: 2-3 days

Description: Implement the core utility endpoints needed by all OpenAI-compatible clients.

Deliverables:

• Implement /v1/models endpoint to list available models
• Add engine lifecycle management in the lifespan handler, following http_server.py in python/sglang/srt/http_server.py
• Add metrics middleware
• Add all openai endpoints, skip implementation for now

Testing:

• Verify proper model listing
• Ensure metrics are correctly exposed