[Frontend] Add request_id to the Request object so they can be controlled better via external load balancers (vllm-project#21009)

Signed-off-by: Kourosh Hakhamaneshi <[email protected]>

Author: kouroshHakha

vllm/entrypoints/openai/protocol.py

@@ -1007,6 +1007,13 @@ class CompletionRequest(OpenAIBaseModel):
             "default: 0). Any priority other than 0 will raise an error "
             "if the served model does not use priority scheduling."),
     )
+    request_id: str = Field(
+        default_factory=lambda: f"{random_uuid()}",
+        description=(
+            "The request_id related to this request. If the caller does "
+            "not set it, a random_uuid will be generated. This id is used "
+            "through out the inference process and return in response."),
+    )
     logits_processors: Optional[LogitsProcessors] = Field(
         default=None,
         description=(
@@ -1251,6 +1258,13 @@ class EmbeddingCompletionRequest(OpenAIBaseModel):
             "default: 0). Any priority other than 0 will raise an error "
             "if the served model does not use priority scheduling."),
     )
+    request_id: str = Field(
+        default_factory=lambda: f"{random_uuid()}",
+        description=(
+            "The request_id related to this request. If the caller does "
+            "not set it, a random_uuid will be generated. This id is used "
+            "through out the inference process and return in response."),
+    )
 
     # --8<-- [end:embedding-extra-params]
 
@@ -1302,6 +1316,13 @@ class EmbeddingChatRequest(OpenAIBaseModel):
             "default: 0). Any priority other than 0 will raise an error "
             "if the served model does not use priority scheduling."),
     )
+    request_id: str = Field(
+        default_factory=lambda: f"{random_uuid()}",
+        description=(
+            "The request_id related to this request. If the caller does "
+            "not set it, a random_uuid will be generated. This id is used "
+            "through out the inference process and return in response."),
+    )
     # --8<-- [end:chat-embedding-extra-params]
 
     @model_validator(mode="before")

vllm/entrypoints/openai/serving_completion.py

@@ -113,7 +113,9 @@ async def create_completion(
             return self.create_error_response(
                 "Echo is unsupported with prompt embeds.")
 
-        request_id = f"cmpl-{self._base_request_id(raw_request)}"
+        request_id = (
+            f"cmpl-"
+            f"{self._base_request_id(raw_request, request.request_id)}")
         created_time = int(time.time())
 
         request_metadata = RequestResponseMetadata(request_id=request_id)

vllm/entrypoints/openai/serving_embedding.py

@@ -163,8 +163,9 @@ async def create_embedding(
         for the API specification. This API mimics the OpenAI Embedding API.
         """
         model_name = self._get_model_name(request.model)
-        request_id = (f"{self.request_id_prefix}-"
-                      f"{self._base_request_id(raw_request)}")
+        request_id = (
+            f"{self.request_id_prefix}-"
+            f"{self._base_request_id(raw_request, request.request_id)}")
 
         ctx = EmbeddingServeContext(
             request=request,