Function connection_manager

fastdeploy/entrypoints/openai/api_server.py:316–327

async context manager for connection manager

()

@asynccontextmanager
async def connection_manager():
    """
    async context manager for connection manager
    """
    try:
        await asyncio.wait_for(connection_semaphore.acquire(), timeout=0.001)
        yield
    except asyncio.TimeoutError:
        api_server_logger.info(f"Reach max request concurrency, semaphore status: {connection_semaphore.status()}")
        raise HTTPException(
            status_code=429, detail=f"Too many requests,current max concurrency is {args.max_concurrency}"
        )


# TODO 传递真实引擎值 通过pid 获取状态