[Core] Expose API endpoint /is_sleeping (#14312)

Signed-off-by: Jun Duan <jun.duan.phd@outlook.com>

vllm/v1/engine/async_llm.py

@@ -407,6 +407,9 @@ class AsyncLLM(EngineClient):
     async def wake_up(self) -> None:
         await self.engine_core.wake_up_async()
 
+    async def is_sleeping(self) -> bool:
+        return await self.engine_core.is_sleeping_async()
+
     async def add_lora(self, lora_request: LoRARequest) -> bool:
         """Load a new LoRA adapter into the engine for future requests."""
         return await self.engine_core.add_lora_async(lora_request)

vllm/v1/engine/core.py

@@ -253,6 +253,9 @@ class EngineCore:
     def wake_up(self):
         self.model_executor.wake_up()
 
+    def is_sleeping(self) -> bool:
+        return self.model_executor.is_sleeping
+
     def execute_dummy_batch(self):
         self.model_executor.collective_rpc("execute_dummy_batch")
 

vllm/v1/engine/core_client.py

@@ -89,6 +89,9 @@ class EngineCoreClient(ABC):
     def wake_up(self) -> None:
         raise NotImplementedError
 
+    def is_sleeping(self) -> bool:
+        raise NotImplementedError
+
     def execute_dummy_batch(self) -> None:
         raise NotImplementedError
 
@@ -128,6 +131,9 @@ class EngineCoreClient(ABC):
     async def wake_up_async(self) -> None:
         raise NotImplementedError
 
+    async def is_sleeping_async(self) -> bool:
+        raise NotImplementedError
+
     async def abort_requests_async(self, request_ids: list[str]) -> None:
         raise NotImplementedError
 
@@ -182,6 +188,9 @@ class InprocClient(EngineCoreClient):
     def wake_up(self) -> None:
         self.engine_core.wake_up()
 
+    def is_sleeping(self) -> bool:
+        return self.engine_core.is_sleeping()
+
     def execute_dummy_batch(self) -> None:
         self.engine_core.execute_dummy_batch()
 
@@ -433,6 +442,9 @@ class SyncMPClient(MPClient):
     def wake_up(self) -> None:
         self._call_utility("wake_up")
 
+    def is_sleeping(self) -> bool:
+        return self._call_utility("is_sleeping")
+
     def execute_dummy_batch(self) -> None:
         self._call_utility("execute_dummy_batch")
 
@@ -523,6 +535,9 @@ class AsyncMPClient(MPClient):
     async def wake_up_async(self) -> None:
         await self._call_utility_async("wake_up")
 
+    async def is_sleeping_async(self) -> bool:
+        return await self._call_utility_async("is_sleeping")
+
     async def execute_dummy_batch_async(self) -> None:
         await self._call_utility_async("execute_dummy_batch")
 

vllm/v1/engine/llm_engine.py

@@ -235,6 +235,9 @@ class LLMEngine:
     def wake_up(self):
         self.engine_core.wake_up()
 
+    def is_sleeping(self) -> bool:
+        return self.engine_core.is_sleeping()
+
     def get_tokenizer_group(
         self,
         group_type: type[_G] = BaseTokenizerGroup,