[router][grpc] Consolidate error messages build in error.rs (#12301)

sgl-router/src/routers/grpc/error.rs

@@ -0,0 +1,143 @@
+//! Centralized error response handling for all routers
+//!
+//! This module provides consistent error responses across OpenAI and gRPC routers,
+//! ensuring all errors follow OpenAI's API error format.
+
+use axum::{
+    http::StatusCode,
+    response::{IntoResponse, Response},
+    Json,
+};
+use serde_json::json;
+use tracing::{error, warn};
+
+/// Create a 500 Internal Server Error response
+///
+/// Use this for unexpected server-side errors, database failures, etc.
+///
+/// # Example
+/// ```ignore
+/// return Err(internal_error("Database connection failed"));
+/// ```
+pub fn internal_error(message: impl Into<String>) -> Response {
+    let msg = message.into();
+    error!("{}", msg);
+    (
+        StatusCode::INTERNAL_SERVER_ERROR,
+        Json(json!({
+            "error": {
+                "message": msg,
+                "type": "internal_error",
+                "code": 500
+            }
+        })),
+    )
+        .into_response()
+}
+
+/// Create a 400 Bad Request response
+///
+/// Use this for invalid request parameters, malformed JSON, validation errors, etc.
+///
+/// # Example
+/// ```ignore
+/// return Err(bad_request("Invalid conversation ID format"));
+/// ```
+pub fn bad_request(message: impl Into<String>) -> Response {
+    let msg = message.into();
+    error!("{}", msg);
+    (
+        StatusCode::BAD_REQUEST,
+        Json(json!({
+            "error": {
+                "message": msg,
+                "type": "invalid_request_error",
+                "code": 400
+            }
+        })),
+    )
+        .into_response()
+}
+
+/// Create a 404 Not Found response
+///
+/// Use this for resources that don't exist (conversations, responses, etc.)
+///
+/// # Example
+/// ```ignore
+/// return Err(not_found(format!("Conversation '{}' not found", id)));
+/// ```
+pub fn not_found(message: impl Into<String>) -> Response {
+    let msg = message.into();
+    warn!("{}", msg);
+    (
+        StatusCode::NOT_FOUND,
+        Json(json!({
+            "error": {
+                "message": msg,
+                "type": "invalid_request_error",
+                "code": 404
+            }
+        })),
+    )
+        .into_response()
+}
+
+/// Create a 503 Service Unavailable response
+///
+/// Use this for temporary service issues like no workers available, rate limiting, etc.
+///
+/// # Example
+/// ```ignore
+/// return Err(service_unavailable("No workers available for this model"));
+/// ```
+pub fn service_unavailable(message: impl Into<String>) -> Response {
+    let msg = message.into();
+    warn!("{}", msg);
+    (
+        StatusCode::SERVICE_UNAVAILABLE,
+        Json(json!({
+            "error": {
+                "message": msg,
+                "type": "service_unavailable",
+                "code": 503
+            }
+        })),
+    )
+        .into_response()
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_internal_error_string() {
+        let response = internal_error("Test error");
+        assert_eq!(response.status(), StatusCode::INTERNAL_SERVER_ERROR);
+    }
+
+    #[test]
+    fn test_internal_error_format() {
+        let response = internal_error(format!("Error: {}", 42));
+        assert_eq!(response.status(), StatusCode::INTERNAL_SERVER_ERROR);
+    }
+
+    #[test]
+    fn test_bad_request() {
+        let response = bad_request("Invalid input");
+        assert_eq!(response.status(), StatusCode::BAD_REQUEST);
+    }
+
+    #[test]
+    fn test_not_found() {
+        let response = not_found("Resource not found");
+        assert_eq!(response.status(), StatusCode::NOT_FOUND);
+    }
+
+    #[test]
+    fn test_service_unavailable() {
+        let response = service_unavailable("No workers");
+        assert_eq!(response.status(), StatusCode::SERVICE_UNAVAILABLE);
+    }
+}

sgl-router/src/routers/grpc/harmony/processor.rs

@@ -18,7 +18,7 @@ use crate::{
     },
     routers::grpc::{
         context::{DispatchMetadata, ExecutionResult},
-        utils,
+        error, utils,
     },
 };
 
@@ -66,7 +66,7 @@ impl HarmonyResponseProcessor {
         // Collect all completed responses (one per choice)
         let all_responses = Self::collect_responses(execution_result).await?;
         if all_responses.is_empty() {
-            return Err(utils::internal_error_static("No responses from server"));
+            return Err(error::internal_error("No responses from server"));
         }
 
         // Build choices by parsing output with HarmonyParserAdapter
@@ -84,7 +84,7 @@ impl HarmonyResponseProcessor {
 
             // Parse Harmony channels with HarmonyParserAdapter
             let mut parser = HarmonyParserAdapter::new().map_err(|e| {
-                utils::internal_error_message(format!("Failed to create Harmony parser: {}", e))
+                error::internal_error(format!("Failed to create Harmony parser: {}", e))
             })?;
 
             // Parse Harmony channels with finish_reason and matched_stop
@@ -94,9 +94,7 @@ impl HarmonyResponseProcessor {
                     complete.finish_reason.clone(),
                     matched_stop.clone(),
                 )
-                .map_err(|e| {
+                .map_err(|e| error::internal_error(format!("Harmony parsing failed: {}", e)))?;
-                    utils::internal_error_message(format!("Harmony parsing failed: {}", e))
-                })?;
 
             // Build response message (assistant)
             let message = ChatCompletionMessage {
@@ -195,17 +193,17 @@ impl HarmonyResponseProcessor {
         // Collect all completed responses
         let all_responses = Self::collect_responses(execution_result).await?;
         if all_responses.is_empty() {
-            return Err(utils::internal_error_static("No responses from server"));
+            return Err(error::internal_error("No responses from server"));
         }
 
         // For Responses API, we only process the first response (n=1)
         let complete = all_responses
             .first()
-            .ok_or_else(|| utils::internal_error_static("No complete response"))?;
+            .ok_or_else(|| error::internal_error("No complete response"))?;
 
         // Parse Harmony channels
         let mut parser = HarmonyParserAdapter::new().map_err(|e| {
-            utils::internal_error_message(format!("Failed to create Harmony parser: {}", e))
+            error::internal_error(format!("Failed to create Harmony parser: {}", e))
         })?;
 
         // Convert matched_stop from proto to JSON
@@ -224,7 +222,7 @@ impl HarmonyResponseProcessor {
                 complete.finish_reason.clone(),
                 matched_stop,
             )
-            .map_err(|e| utils::internal_error_message(format!("Harmony parsing failed: {}", e)))?;
+            .map_err(|e| error::internal_error(format!("Harmony parsing failed: {}", e)))?;
 
         // VALIDATION: Check if model incorrectly generated Tool role messages
         // This happens when the model copies the format of tool result messages

sgl-router/src/routers/grpc/harmony/responses.rs

@@ -61,10 +61,10 @@ use crate::{
     routers::{
         grpc::{
             context::SharedComponents,
+            error,
             harmony::processor::ResponsesIterationResult,
             pipeline::RequestPipeline,
             responses::streaming::{OutputItemType, ResponseStreamEventEmitter},
-            utils,
         },
         openai::mcp::ensure_request_mcp_client,
     },
@@ -285,7 +285,7 @@ pub async fn serve_harmony_responses(
 
         // Safety check: prevent infinite loops
         if iteration_count > MAX_TOOL_ITERATIONS {
-            return Err(utils::internal_error_message(format!(
+            return Err(error::internal_error(format!(
                 "Maximum tool iterations ({}) exceeded",
                 MAX_TOOL_ITERATIONS
             )));
@@ -333,7 +333,7 @@ pub async fn serve_harmony_responses(
                     execute_mcp_tools(&ctx.mcp_manager, &tool_calls, tracking).await?
                 } else {
                     // Should never happen (we only get tool_calls when has_mcp_tools=true)
-                    return Err(utils::internal_error_static(
+                    return Err(error::internal_error(
                         "Tool calls found but MCP tracking not initialized",
                     ));
                 };
@@ -734,7 +734,7 @@ async fn execute_mcp_tools(
         // Parse tool arguments from JSON string
         let args_str = tool_call.function.arguments.as_deref().unwrap_or("{}");
         let args: JsonValue = serde_json::from_str(args_str).map_err(|e| {
-            utils::internal_error_message(format!(
+            error::internal_error(format!(
                 "Invalid tool arguments JSON for tool '{}': {}",
                 tool_call.function.name, e
             ))
@@ -1111,7 +1111,7 @@ async fn load_previous_messages(
         .get_response_chain(&prev_id, None)
         .await
         .map_err(|e| {
-            utils::internal_error_message(format!(
+            error::internal_error(format!(
                 "Failed to load previous response chain for {}: {}",
                 prev_id_str, e
             ))

sgl-router/src/routers/grpc/harmony/stages/preparation.rs

@@ -13,6 +13,7 @@ use crate::{
     },
     routers::grpc::{
         context::{PreparationOutput, RequestContext, RequestType},
+        error,
         stages::PipelineStage,
         utils,
     },
@@ -56,7 +57,7 @@ impl PipelineStage for HarmonyPreparationStage {
             let request_arc = ctx.responses_request_arc();
             self.prepare_responses(ctx, &request_arc).await?;
         } else {
-            return Err(utils::bad_request_error(
+            return Err(error::bad_request(
                 "Only Chat and Responses requests supported in Harmony pipeline".to_string(),
             ));
         }
@@ -78,7 +79,7 @@ impl HarmonyPreparationStage {
     ) -> Result<Option<Response>, Response> {
         // Validate - reject logprobs
         if request.logprobs {
-            return Err(utils::bad_request_error(
+            return Err(error::bad_request(
                 "logprobs are not supported for Harmony models".to_string(),
             ));
         }
@@ -97,7 +98,7 @@ impl HarmonyPreparationStage {
         let build_output = self
             .builder
             .build_from_chat(&body_ref)
-            .map_err(|e| utils::bad_request_error(format!("Harmony build failed: {}", e)))?;
+            .map_err(|e| error::bad_request(format!("Harmony build failed: {}", e)))?;
 
         // Step 4: Store results
         ctx.state.preparation = Some(PreparationOutput {
@@ -132,7 +133,7 @@ impl HarmonyPreparationStage {
         let build_output = self
             .builder
             .build_from_responses(request)
-            .map_err(|e| utils::bad_request_error(format!("Harmony build failed: {}", e)))?;
+            .map_err(|e| error::bad_request(format!("Harmony build failed: {}", e)))?;
 
         // Store results in preparation output
         ctx.state.preparation = Some(PreparationOutput {
@@ -202,7 +203,7 @@ impl HarmonyPreparationStage {
 
         // Validate specific function exists
         if specific_function.is_some() && tools_to_use.is_empty() {
-            return Err(Box::new(utils::bad_request_error(format!(
+            return Err(Box::new(error::bad_request(format!(
                 "Tool '{}' not found in tools list",
                 specific_function.unwrap()
             ))));
@@ -236,7 +237,7 @@ impl HarmonyPreparationStage {
         });
 
         serde_json::to_string(&structural_tag).map_err(|e| {
-            Box::new(utils::internal_error_message(format!(
+            Box::new(error::internal_error(format!(
                 "Failed to serialize structural tag: {}",
                 e
             )))

sgl-router/src/routers/grpc/harmony/stages/request_building.rs

@@ -13,8 +13,8 @@ use crate::{
     grpc_client::proto::{DisaggregatedParams, GenerateRequest},
     routers::grpc::{
         context::{ClientSelection, RequestContext, RequestType, WorkerSelection},
+        error,
         stages::PipelineStage,
-        utils,
     },
 };
 
@@ -69,14 +69,14 @@ impl PipelineStage for HarmonyRequestBuildingStage {
             .state
             .preparation
             .as_ref()
-            .ok_or_else(|| utils::internal_error_static("Preparation not completed"))?;
+            .ok_or_else(|| error::internal_error("Preparation not completed"))?;
 
         // Get clients
         let clients = ctx
             .state
             .clients
             .as_ref()
-            .ok_or_else(|| utils::internal_error_static("Client acquisition not completed"))?;
+            .ok_or_else(|| error::internal_error("Client acquisition not completed"))?;
         let builder_client = match clients {
             ClientSelection::Single { client } => client,
             ClientSelection::Dual { prefill, .. } => prefill,
@@ -87,7 +87,7 @@ impl PipelineStage for HarmonyRequestBuildingStage {
             RequestType::Chat(_) => format!("chatcmpl-{}", Uuid::new_v4()),
             RequestType::Responses(_) => format!("responses-{}", Uuid::new_v4()),
             RequestType::Generate(_) => {
-                return Err(utils::bad_request_error(
+                return Err(error::bad_request(
                     "Generate requests are not supported with Harmony models".to_string(),
                 ));
             }
@@ -111,9 +111,7 @@ impl PipelineStage for HarmonyRequestBuildingStage {
                         None,
                         prep.tool_constraints.clone(),
                     )
-                    .map_err(|e| {
+                    .map_err(|e| error::bad_request(format!("Invalid request parameters: {}", e)))?
-                        utils::bad_request_error(format!("Invalid request parameters: {}", e))
-                    })?
             }
             RequestType::Responses(request) => builder_client
                 .build_generate_request_from_responses(
@@ -123,9 +121,7 @@ impl PipelineStage for HarmonyRequestBuildingStage {
                     prep.token_ids.clone(),
                     prep.harmony_stop_ids.clone(),
                 )
-                .map_err(|e| {
+                .map_err(|e| error::bad_request(format!("Invalid request parameters: {}", e)))?,
-                    utils::bad_request_error(format!("Invalid request parameters: {}", e))
-                })?,
             _ => unreachable!(),
         };
 

sgl-router/src/routers/grpc/harmony/stages/response_processing.rs

@@ -8,8 +8,8 @@ use axum::response::Response;
 use super::super::{HarmonyResponseProcessor, HarmonyStreamingProcessor};
 use crate::routers::grpc::{
     context::{FinalResponse, RequestContext, RequestType},
+    error,
     stages::PipelineStage,
-    utils,
 };
 
 /// Harmony Response Processing stage: Parse and format Harmony responses
@@ -51,14 +51,14 @@ impl PipelineStage for HarmonyResponseProcessingStage {
                     .response
                     .execution_result
                     .take()
-                    .ok_or_else(|| utils::internal_error_static("No execution result"))?;
+                    .ok_or_else(|| error::internal_error("No execution result"))?;
 
                 let dispatch = ctx
                     .state
                     .dispatch
                     .as_ref()
                     .cloned()
-                    .ok_or_else(|| utils::internal_error_static("Dispatch metadata not set"))?;
+                    .ok_or_else(|| error::internal_error("Dispatch metadata not set"))?;
 
                 // For streaming, delegate to streaming processor and return SSE response
                 if is_streaming {
@@ -97,14 +97,14 @@ impl PipelineStage for HarmonyResponseProcessingStage {
                     .response
                     .execution_result
                     .take()
-                    .ok_or_else(|| utils::internal_error_static("No execution result"))?;
+                    .ok_or_else(|| error::internal_error("No execution result"))?;
 
                 let dispatch = ctx
                     .state
                     .dispatch
                     .as_ref()
                     .cloned()
-                    .ok_or_else(|| utils::internal_error_static("Dispatch metadata not set"))?;
+                    .ok_or_else(|| error::internal_error("Dispatch metadata not set"))?;
 
                 let responses_request = ctx.responses_request_arc();
                 let iteration_result = self
@@ -115,7 +115,7 @@ impl PipelineStage for HarmonyResponseProcessingStage {
                 ctx.state.response.responses_iteration_result = Some(iteration_result);
                 Ok(None)
             }
-            RequestType::Generate(_) => Err(utils::internal_error_static(
+            RequestType::Generate(_) => Err(error::internal_error(
                 "Generate requests not supported in Harmony pipeline",
             )),
         }

sgl-router/src/routers/grpc/mod.rs

@@ -3,6 +3,7 @@
 use crate::{grpc_client::proto, protocols::common::StringOrArray};
 
 pub mod context;
+pub mod error;
 pub mod harmony;
 pub mod pd_router;
 pub mod pipeline;

sgl-router/src/routers/grpc/pipeline.rs

@@ -11,7 +11,7 @@ use tracing::{debug, error};
 
 // Import all stage types from the stages module
 use super::stages::*;
-use super::{context::*, harmony, processing, responses::BackgroundTaskInfo, streaming, utils};
+use super::{context::*, error, harmony, processing, responses::BackgroundTaskInfo, streaming};
 use crate::{
     core::WorkerRegistry,
     policies::PolicyRegistry,
@@ -228,9 +228,9 @@ impl RequestPipeline {
         match ctx.state.response.final_response {
             Some(FinalResponse::Chat(response)) => axum::Json(response).into_response(),
             Some(FinalResponse::Generate(_)) => {
-                utils::internal_error_static("Internal error: wrong response type")
+                error::internal_error("Internal error: wrong response type")
             }
-            None => utils::internal_error_static("No response produced"),
+            None => error::internal_error("No response produced"),
         }
     }
 
@@ -272,9 +272,9 @@ impl RequestPipeline {
         match ctx.state.response.final_response {
             Some(FinalResponse::Generate(response)) => axum::Json(response).into_response(),
             Some(FinalResponse::Chat(_)) => {
-                utils::internal_error_static("Internal error: wrong response type")
+                error::internal_error("Internal error: wrong response type")
             }
-            None => utils::internal_error_static("No response produced"),
+            None => error::internal_error("No response produced"),
         }
     }
 
@@ -303,7 +303,7 @@ impl RequestPipeline {
             match stage.execute(&mut ctx).await {
                 Ok(Some(_response)) => {
                     // Streaming not supported for responses sync mode
-                    return Err(utils::bad_request_error(
+                    return Err(error::bad_request(
                         "Streaming is not supported in this context".to_string(),
                     ));
                 }
@@ -360,10 +360,10 @@ impl RequestPipeline {
         // Extract final response
         match ctx.state.response.final_response {
             Some(FinalResponse::Chat(response)) => Ok(response),
-            Some(FinalResponse::Generate(_)) => Err(utils::internal_error_static(
+            Some(FinalResponse::Generate(_)) => {
-                "Internal error: wrong response type",
+                Err(error::internal_error("Internal error: wrong response type"))
-            )),
+            }
-            None => Err(utils::internal_error_static("No response produced")),
+            None => Err(error::internal_error("No response produced")),
         }
     }
 
@@ -384,7 +384,7 @@ impl RequestPipeline {
         _model_id: Option<String>,
         _components: Arc<SharedComponents>,
     ) -> Response {
-        utils::internal_error_static("Responses API execution not yet implemented")
+        error::internal_error("Responses API execution not yet implemented")
     }
 
     /// Execute Harmony Responses API request through all pipeline stages
@@ -451,7 +451,7 @@ impl RequestPipeline {
             .responses_iteration_result
             .take()
             .ok_or_else(|| {
-                utils::internal_error_static("No ResponsesIterationResult produced by pipeline")
+                error::internal_error("No ResponsesIterationResult produced by pipeline")
             })
     }
 
@@ -501,6 +501,6 @@ impl RequestPipeline {
             .response
             .execution_result
             .take()
-            .ok_or_else(|| utils::internal_error_static("No ExecutionResult produced by pipeline"))
+            .ok_or_else(|| error::internal_error("No ExecutionResult produced by pipeline"))
     }
 }

sgl-router/src/routers/grpc/processing.rs

@@ -11,7 +11,7 @@ use tracing::error;
 
 use super::{
     context::{DispatchMetadata, ExecutionResult},
-    utils,
+    error, utils,
 };
 use crate::{
     grpc_client::proto,
@@ -104,7 +104,7 @@ impl ResponseProcessor {
         };
 
         if all_responses.is_empty() {
-            return Err(utils::internal_error_static("No responses from server"));
+            return Err(error::internal_error("No responses from server"));
         }
 
         Ok(all_responses)
@@ -332,7 +332,7 @@ impl ResponseProcessor {
             {
                 Ok(choice) => choices.push(choice),
                 Err(e) => {
-                    return Err(utils::internal_error_message(format!(
+                    return Err(error::internal_error(format!(
                         "Failed to process choice {}: {}",
                         index, e
                     )));
@@ -447,7 +447,7 @@ impl ResponseProcessor {
             let outputs = match stop_decoder.process_tokens(&complete.output_ids) {
                 Ok(outputs) => outputs,
                 Err(e) => {
-                    return Err(utils::internal_error_message(format!(
+                    return Err(error::internal_error(format!(
                         "Failed to process tokens: {}",
                         e
                     )))

sgl-router/src/routers/grpc/responses/handlers.rs

@@ -67,7 +67,10 @@ use crate::{
             ResponseStatus, ResponsesRequest, ResponsesResponse, ResponsesUsage,
         },
     },
-    routers::openai::{conversations::persist_conversation_items, mcp::ensure_request_mcp_client},
+    routers::{
+        grpc::error,
+        openai::{conversations::persist_conversation_items, mcp::ensure_request_mcp_client},
+    },
 };
 
 // ============================================================================
@@ -863,11 +866,9 @@ async fn execute_without_mcp(
     model_id: Option<String>,
     response_id: Option<String>,
 ) -> Result<ResponsesResponse, Response> {
-    use crate::routers::grpc::utils;
-
     // Convert ResponsesRequest → ChatCompletionRequest
     let chat_request = conversions::responses_to_chat(modified_request)
-        .map_err(|e| utils::bad_request_error(format!("Failed to convert request: {}", e)))?;
+        .map_err(|e| error::bad_request(format!("Failed to convert request: {}", e)))?;
 
     // Execute chat pipeline (errors already have proper HTTP status codes)
     let chat_response = ctx
@@ -883,9 +884,8 @@ async fn execute_without_mcp(
         .await?; // Preserve the Response error as-is
 
     // Convert ChatCompletionResponse → ResponsesResponse
-    conversions::chat_to_responses(&chat_response, original_request, response_id).map_err(|e| {
+    conversions::chat_to_responses(&chat_response, original_request, response_id)
-        utils::internal_error_message(format!("Failed to convert to responses format: {}", e))
+        .map_err(|e| error::internal_error(format!("Failed to convert to responses format: {}", e)))
-    })
 }
 
 /// Load conversation history and response chains, returning modified request
@@ -962,15 +962,10 @@ async fn load_conversation_history(
             .conversation_storage
             .get_conversation(&conv_id)
             .await
-            .map_err(|e| {
+            .map_err(|e| error::internal_error(format!("Failed to check conversation: {}", e)))?;
-                crate::routers::grpc::utils::internal_error_message(format!(
-                    "Failed to check conversation: {}",
-                    e
-                ))
-            })?;
 
         if conversation.is_none() {
-            return Err(crate::routers::grpc::utils::bad_request_error(format!(
+            return Err(error::not_found(format!(
                 "Conversation '{}' not found. Please create the conversation first using the conversations API.",
                 conv_id_str
             )));

sgl-router/src/routers/grpc/responses/tool_loop.rs

@@ -19,6 +19,7 @@ use tracing::{debug, warn};
 use uuid::Uuid;
 
 use super::{
+    super::error,
     conversions,
     streaming::{OutputItemType, ResponseStreamEventEmitter},
 };
@@ -247,12 +248,8 @@ pub(super) async fn execute_tool_loop(
 
     loop {
         // Convert to chat request
-        let mut chat_request = conversions::responses_to_chat(&current_request).map_err(|e| {
+        let mut chat_request = conversions::responses_to_chat(&current_request)
-            crate::routers::grpc::utils::bad_request_error(format!(
+            .map_err(|e| error::bad_request(format!("Failed to convert request: {}", e)))?;
-                "Failed to convert request: {}",
-                e
-            ))
-        })?;
 
         // Add MCP tools to chat request so LLM knows about them
         chat_request.tools = Some(chat_tools.clone());
@@ -301,10 +298,7 @@ pub(super) async fn execute_tool_loop(
                     response_id.clone(),
                 )
                 .map_err(|e| {
-                    crate::routers::grpc::utils::internal_error_message(format!(
+                    error::internal_error(format!("Failed to convert to responses format: {}", e))
-                        "Failed to convert to responses format: {}",
-                        e
-                    ))
                 })?;
 
                 // Mark as completed but with incomplete details
@@ -423,10 +417,7 @@ pub(super) async fn execute_tool_loop(
                 response_id.clone(),
             )
             .map_err(|e| {
-                crate::routers::grpc::utils::internal_error_message(format!(
+                error::internal_error(format!("Failed to convert to responses format: {}", e))
-                    "Failed to convert to responses format: {}",
-                    e
-                ))
             })?;
 
             // Inject MCP metadata into output

sgl-router/src/routers/grpc/stages/client_acquisition.rs

@@ -6,7 +6,7 @@ use axum::response::Response;
 use super::PipelineStage;
 use crate::routers::grpc::{
     context::{ClientSelection, RequestContext, WorkerSelection},
-    utils,
+    error, utils,
 };
 
 /// Client acquisition stage: Get gRPC clients from selected workers
@@ -19,7 +19,7 @@ impl PipelineStage for ClientAcquisitionStage {
             .state
             .workers
             .as_ref()
-            .ok_or_else(|| utils::internal_error_static("Worker selection not completed"))?;
+            .ok_or_else(|| error::internal_error("Worker selection not completed"))?;
 
         let clients = match workers {
             WorkerSelection::Single { worker } => {

sgl-router/src/routers/grpc/stages/dispatch_metadata.rs

@@ -8,7 +8,7 @@ use axum::response::Response;
 use super::PipelineStage;
 use crate::routers::grpc::{
     context::{DispatchMetadata, RequestContext, RequestType, WorkerSelection},
-    utils,
+    error,
 };
 
 /// Dispatch metadata stage: Prepare metadata for dispatch
@@ -21,7 +21,7 @@ impl PipelineStage for DispatchMetadataStage {
             .state
             .proto_request
             .as_ref()
-            .ok_or_else(|| utils::internal_error_static("Proto request not built"))?;
+            .ok_or_else(|| error::internal_error("Proto request not built"))?;
 
         let request_id = proto_request.request_id.clone();
         let model = match &ctx.input.request_type {

sgl-router/src/routers/grpc/stages/preparation.rs

@@ -10,7 +10,7 @@ use crate::{
     protocols::{chat::ChatCompletionRequest, common::InputIds, generate::GenerateRequest},
     routers::grpc::{
         context::{PreparationOutput, RequestContext, RequestType},
-        utils,
+        error, utils,
     },
     tokenizer::traits::Tokenizer,
 };
@@ -56,7 +56,7 @@ impl PreparationStage {
             match utils::process_chat_messages(&body_ref, &*ctx.components.tokenizer) {
                 Ok(msgs) => msgs,
                 Err(e) => {
-                    return Err(utils::bad_request_error(e));
+                    return Err(error::bad_request(e));
                 }
             };
 
@@ -64,10 +64,7 @@ impl PreparationStage {
         let encoding = match ctx.components.tokenizer.encode(&processed_messages.text) {
             Ok(encoding) => encoding,
             Err(e) => {
-                return Err(utils::internal_error_message(format!(
+                return Err(error::internal_error(format!("Tokenization failed: {}", e)));
-                    "Tokenization failed: {}",
-                    e
-                )));
             }
         };
 
@@ -75,9 +72,8 @@ impl PreparationStage {
 
         // Step 4: Build tool constraints if needed
         let tool_call_constraint = if let Some(tools) = body_ref.tools.as_ref() {
-            utils::generate_tool_constraints(tools, &request.tool_choice, &request.model).map_err(
+            utils::generate_tool_constraints(tools, &request.tool_choice, &request.model)
-                |e| utils::bad_request_error(format!("Invalid tool configuration: {}", e)),
+                .map_err(|e| error::bad_request(format!("Invalid tool configuration: {}", e)))?
-            )?
         } else {
             None
         };
@@ -124,7 +120,7 @@ impl PreparationStage {
         let (original_text, token_ids) = match self.resolve_generate_input(ctx, request) {
             Ok(res) => res,
             Err(msg) => {
-                return Err(utils::bad_request_error(msg));
+                return Err(error::bad_request(msg));
             }
         };
 

sgl-router/src/routers/grpc/stages/request_building.rs

@@ -15,7 +15,7 @@ use crate::{
     grpc_client::proto,
     routers::grpc::{
         context::{ClientSelection, RequestContext, RequestType, WorkerSelection},
-        utils,
+        error,
     },
 };
 
@@ -37,13 +37,13 @@ impl PipelineStage for RequestBuildingStage {
             .state
             .preparation
             .as_ref()
-            .ok_or_else(|| utils::internal_error_static("Preparation not completed"))?;
+            .ok_or_else(|| error::internal_error("Preparation not completed"))?;
 
         let clients = ctx
             .state
             .clients
             .as_ref()
-            .ok_or_else(|| utils::internal_error_static("Client acquisition not completed"))?;
+            .ok_or_else(|| error::internal_error("Client acquisition not completed"))?;
 
         // Get client for building request (use prefill client if PD mode)
         let builder_client = match clients {
@@ -69,9 +69,7 @@ impl PipelineStage for RequestBuildingStage {
                             .clone(),
                         prep.tool_constraints.clone(),
                     )
-                    .map_err(|e| {
+                    .map_err(|e| error::bad_request(format!("Invalid request parameters: {}", e)))?
-                        utils::bad_request_error(format!("Invalid request parameters: {}", e))
-                    })?
             }
             RequestType::Generate(request) => {
                 let request_id = request
@@ -86,7 +84,7 @@ impl PipelineStage for RequestBuildingStage {
                         prep.original_text.clone(),
                         prep.token_ids.clone(),
                     )
-                    .map_err(utils::bad_request_error)?
+                    .map_err(error::bad_request)?
             }
             RequestType::Responses(_request) => {
                 // Responses API builds request during the MCP loop

sgl-router/src/routers/grpc/stages/request_execution.rs

@@ -5,13 +5,15 @@ use axum::response::Response;
 
 use super::PipelineStage;
 use crate::{
-    grpc_client::proto,
+    grpc_client::{proto, sglang_scheduler::AbortOnDropStream},
     routers::grpc::{
         context::{ClientSelection, ExecutionResult, RequestContext},
-        utils,
+        error,
     },
 };
 
+type StreamResult = Result<AbortOnDropStream, Box<dyn std::error::Error + Send + Sync>>;
+
 /// Request execution stage: Execute gRPC requests (single or dual dispatch)
 pub struct RequestExecutionStage {
     mode: ExecutionMode,
@@ -37,13 +39,13 @@ impl PipelineStage for RequestExecutionStage {
             .state
             .proto_request
             .take()
-            .ok_or_else(|| utils::internal_error_static("Proto request not built"))?;
+            .ok_or_else(|| error::internal_error("Proto request not built"))?;
 
         let clients = ctx
             .state
             .clients
             .as_mut()
-            .ok_or_else(|| utils::internal_error_static("Client acquisition not completed"))?;
+            .ok_or_else(|| error::internal_error("Client acquisition not completed"))?;
 
         let result = match self.mode {
             ExecutionMode::Single => self.execute_single(proto_request, clients).await?,
@@ -70,11 +72,12 @@ impl RequestExecutionStage {
     ) -> Result<ExecutionResult, Response> {
         let client = clients
             .single_mut()
-            .ok_or_else(|| utils::internal_error_static("Expected single client but got dual"))?;
+            .ok_or_else(|| error::internal_error("Expected single client but got dual"))?;
 
-        let stream = client.generate(proto_request).await.map_err(|e| {
+        let stream = client
-            utils::internal_error_message(format!("Failed to start generation: {}", e))
+            .generate(proto_request)
-        })?;
+            .await
+            .map_err(|e| error::internal_error(format!("Failed to start generation: {}", e)))?;
 
         Ok(ExecutionResult::Single { stream })
     }
@@ -86,12 +89,12 @@ impl RequestExecutionStage {
     ) -> Result<ExecutionResult, Response> {
         let (prefill_client, decode_client) = clients
             .dual_mut()
-            .ok_or_else(|| utils::internal_error_static("Expected dual clients but got single"))?;
+            .ok_or_else(|| error::internal_error("Expected dual clients but got single"))?;
 
         let prefill_request = proto_request.clone();
         let decode_request = proto_request;
 
-        let (prefill_result, decode_result) = tokio::join!(
+        let (prefill_result, decode_result): (StreamResult, StreamResult) = tokio::join!(
             prefill_client.generate(prefill_request),
             decode_client.generate(decode_request)
         );
@@ -100,7 +103,7 @@ impl RequestExecutionStage {
         let prefill_stream = match prefill_result {
             Ok(s) => s,
             Err(e) => {
-                return Err(utils::internal_error_message(format!(
+                return Err(error::internal_error(format!(
                     "Prefill worker failed to start: {}",
                     e
                 )));
@@ -111,7 +114,7 @@ impl RequestExecutionStage {
         let decode_stream = match decode_result {
             Ok(s) => s,
             Err(e) => {
-                return Err(utils::internal_error_message(format!(
+                return Err(error::internal_error(format!(
                     "Decode worker failed to start: {}",
                     e
                 )));

sgl-router/src/routers/grpc/stages/response_processing.rs

@@ -11,7 +11,7 @@ use axum::response::Response;
 use super::PipelineStage;
 use crate::routers::grpc::{
     context::{FinalResponse, RequestContext, RequestType},
-    processing, streaming, utils,
+    error, processing, streaming,
 };
 
 /// Response processing stage: Handles both streaming and non-streaming responses
@@ -42,7 +42,7 @@ impl PipelineStage for ResponseProcessingStage {
         match &ctx.input.request_type {
             RequestType::Chat(_) => self.process_chat_response(ctx).await,
             RequestType::Generate(_) => self.process_generate_response(ctx).await,
-            RequestType::Responses(_) => Err(utils::bad_request_error(
+            RequestType::Responses(_) => Err(error::bad_request(
                 "Responses API processing must be handled by responses handler".to_string(),
             )),
         }
@@ -66,14 +66,14 @@ impl ResponseProcessingStage {
             .response
             .execution_result
             .take()
-            .ok_or_else(|| utils::internal_error_static("No execution result"))?;
+            .ok_or_else(|| error::internal_error("No execution result"))?;
 
         // Get dispatch metadata (needed by both streaming and non-streaming)
         let dispatch = ctx
             .state
             .dispatch
             .as_ref()
-            .ok_or_else(|| utils::internal_error_static("Dispatch metadata not set"))?
+            .ok_or_else(|| error::internal_error("Dispatch metadata not set"))?
             .clone();
 
         if is_streaming {
@@ -100,7 +100,7 @@ impl ResponseProcessingStage {
             .response
             .stop_decoder
             .as_mut()
-            .ok_or_else(|| utils::internal_error_static("Stop decoder not initialized"))?;
+            .ok_or_else(|| error::internal_error("Stop decoder not initialized"))?;
 
         let response = self
             .processor
@@ -132,14 +132,14 @@ impl ResponseProcessingStage {
             .response
             .execution_result
             .take()
-            .ok_or_else(|| utils::internal_error_static("No execution result"))?;
+            .ok_or_else(|| error::internal_error("No execution result"))?;
 
         // Get dispatch metadata (needed by both streaming and non-streaming)
         let dispatch = ctx
             .state
             .dispatch
             .as_ref()
-            .ok_or_else(|| utils::internal_error_static("Dispatch metadata not set"))?
+            .ok_or_else(|| error::internal_error("Dispatch metadata not set"))?
             .clone();
 
         if is_streaming {
@@ -162,7 +162,7 @@ impl ResponseProcessingStage {
             .response
             .stop_decoder
             .as_mut()
-            .ok_or_else(|| utils::internal_error_static("Stop decoder not initialized"))?;
+            .ok_or_else(|| error::internal_error("Stop decoder not initialized"))?;
 
         let result_array = self
             .processor

sgl-router/src/routers/grpc/stages/worker_selection.rs

@@ -12,7 +12,7 @@ use crate::{
     policies::PolicyRegistry,
     routers::grpc::{
         context::{RequestContext, WorkerSelection},
-        utils,
+        error,
     },
 };
 
@@ -51,7 +51,7 @@ impl PipelineStage for WorkerSelectionStage {
             .state
             .preparation
             .as_ref()
-            .ok_or_else(|| utils::internal_error_static("Preparation stage not completed"))?;
+            .ok_or_else(|| error::internal_error("Preparation stage not completed"))?;
 
         // For Harmony, use selection_text produced during Harmony encoding
         // Otherwise, use original_text from regular preparation
@@ -66,7 +66,7 @@ impl PipelineStage for WorkerSelectionStage {
                 match self.select_single_worker(ctx.input.model_id.as_deref(), text) {
                     Some(w) => WorkerSelection::Single { worker: w },
                     None => {
-                        return Err(utils::service_unavailable_error(format!(
+                        return Err(error::service_unavailable(format!(
                             "No available workers for model: {:?}",
                             ctx.input.model_id
                         )));
@@ -77,7 +77,7 @@ impl PipelineStage for WorkerSelectionStage {
                 match self.select_pd_pair(ctx.input.model_id.as_deref(), text) {
                     Some((prefill, decode)) => WorkerSelection::Dual { prefill, decode },
                     None => {
-                        return Err(utils::service_unavailable_error(format!(
+                        return Err(error::service_unavailable(format!(
                             "No available PD worker pairs for model: {:?}",
                             ctx.input.model_id
                         )));

sgl-router/src/routers/grpc/utils.rs

@@ -2,17 +2,13 @@
 
 use std::{collections::HashMap, sync::Arc};
 
-use axum::{
+use axum::response::Response;
-    http::StatusCode,
-    response::{IntoResponse, Response},
-    Json,
-};
 use futures::StreamExt;
 use serde_json::{json, Map, Value};
 use tracing::{error, warn};
 use uuid::Uuid;
 
-use super::ProcessedMessages;
+use super::{error, ProcessedMessages};
 pub use crate::tokenizer::StopSequenceDecoder;
 use crate::{
     core::Worker,
@@ -40,8 +36,8 @@ pub async fn get_grpc_client_from_worker(
     let client_arc = worker
         .get_grpc_client()
         .await
-        .map_err(|e| internal_error_message(format!("Failed to get gRPC client: {}", e)))?
+        .map_err(|e| error::internal_error(format!("Failed to get gRPC client: {}", e)))?
-        .ok_or_else(|| internal_error_static("Selected worker is not configured for gRPC"))?;
+        .ok_or_else(|| error::internal_error("Selected worker is not configured for gRPC"))?;
 
     Ok((*client_arc).clone())
 }
@@ -433,67 +429,6 @@ pub fn process_chat_messages(
     })
 }
 
-/// Error response helpers (shared between regular and PD routers)
-pub fn internal_error_static(msg: &'static str) -> Response {
-    error!("{}", msg);
-    (
-        StatusCode::INTERNAL_SERVER_ERROR,
-        Json(json!({
-            "error": {
-                "message": msg,
-                "type": "internal_error",
-                "code": 500
-            }
-        })),
-    )
-        .into_response()
-}
-
-pub fn internal_error_message(message: String) -> Response {
-    error!("{}", message);
-    (
-        StatusCode::INTERNAL_SERVER_ERROR,
-        Json(json!({
-            "error": {
-                "message": message,
-                "type": "internal_error",
-                "code": 500
-            }
-        })),
-    )
-        .into_response()
-}
-
-pub fn bad_request_error(message: String) -> Response {
-    error!("{}", message);
-    (
-        StatusCode::BAD_REQUEST,
-        Json(json!({
-            "error": {
-                "message": message,
-                "type": "invalid_request_error",
-                "code": 400
-            }
-        })),
-    )
-        .into_response()
-}
-
-pub fn service_unavailable_error(message: String) -> Response {
-    warn!("{}", message);
-    (
-        StatusCode::SERVICE_UNAVAILABLE,
-        Json(json!({
-            "error": {
-                "message": message,
-                "type": "service_unavailable",
-                "code": 503
-            }
-        })),
-    )
-        .into_response()
-}
-
 /// Create a StopSequenceDecoder from stop parameters
 pub fn create_stop_decoder(
     tokenizer: &Arc<dyn Tokenizer>,
@@ -646,7 +581,7 @@ pub async fn collect_stream_responses(
                     Some(Error(err)) => {
                         error!("{} error: {}", worker_name, err.message);
                         // Don't mark as completed - let Drop send abort for error cases
-                        return Err(internal_error_message(format!(
+                        return Err(error::internal_error(format!(
                             "{} generation failed: {}",
                             worker_name, err.message
                         )));
@@ -662,7 +597,7 @@ pub async fn collect_stream_responses(
             Err(e) => {
                 error!("{} stream error: {:?}", worker_name, e);
                 // Don't mark as completed - let Drop send abort for error cases
-                return Err(internal_error_message(format!(
+                return Err(error::internal_error(format!(
                     "{} stream failed: {}",
                     worker_name, e
                 )));