Add retry logic to API handlers with @withRetry decorator (RooCodeInc#3596)

Author: ViezeVingertjes

src/api/providers/asksage.ts

@@ -9,6 +9,7 @@ import {
 	askSageDefaultURL,
 } from "@shared/api"
 import { ApiStream } from "../transform/stream"
+import { withRetry } from "../retry"
 
 type AskSageRequest = {
 	system_prompt: string
@@ -45,6 +46,7 @@ export class AskSageHandler implements ApiHandler {
 		}
 	}
 
+	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		try {
 			const model = this.getModel()

src/api/providers/cline.ts

@@ -6,6 +6,7 @@ import { createOpenRouterStream } from "../transform/openrouter-stream"
 import { ApiStream, ApiStreamUsageChunk } from "../transform/stream"
 import axios from "axios"
 import { OpenRouterErrorResponse } from "./types"
+import { withRetry } from "../retry"
 
 export class ClineHandler implements ApiHandler {
 	private options: ApiHandlerOptions
@@ -25,6 +26,7 @@ export class ClineHandler implements ApiHandler {
 		})
 	}
 
+	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		this.lastGenerationId = undefined
 

src/api/providers/doubao.ts

@@ -4,6 +4,7 @@ import { Anthropic } from "@anthropic-ai/sdk"
 import OpenAI from "openai"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"
+import { withRetry } from "../retry"
 
 export class DoubaoHandler implements ApiHandler {
 	private options: ApiHandlerOptions
@@ -28,6 +29,7 @@ export class DoubaoHandler implements ApiHandler {
 		}
 	}
 
+	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		const model = this.getModel()
 		let openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [

src/api/providers/gemini.ts

@@ -171,8 +171,22 @@ export class GeminiHandler implements ApiHandler {
 
 				// Gemini doesn't include status codes in their errors
 				// https://github.com/googleapis/js-genai/blob/61f7f27b866c74333ca6331883882489bcb708b9/src/_api_client.ts#L569
-				if (error.name === "ClientError" && error.message.includes("got status: 429 Too Many Requests.")) {
-					;(error as any).status = 429
+				const rateLimitPatterns = [
+					/got status: 429/i,
+					/429 Too Many Requests/i,
+					/rate limit exceeded/i,
+					/too many requests/i,
+				]
+
+				const isRateLimit =
+					error.name === "ClientError" && rateLimitPatterns.some((pattern) => pattern.test(error.message))
+
+				if (isRateLimit) {
+					const rateLimitError = Object.assign(new Error(error.message), {
+						...error,
+						status: 429,
+					})
+					throw rateLimitError
 				}
 			} else {
 				apiError = String(error)

src/api/providers/litellm.ts

@@ -4,6 +4,7 @@ import { ApiHandlerOptions, liteLlmDefaultModelId, liteLlmModelInfoSaneDefaults
 import { ApiHandler } from ".."
 import { ApiStream } from "../transform/stream"
 import { convertToOpenAiMessages } from "../transform/openai-format"
+import { withRetry } from "../retry"
 
 export class LiteLlmHandler implements ApiHandler {
 	private options: ApiHandlerOptions
@@ -51,6 +52,7 @@ export class LiteLlmHandler implements ApiHandler {
 		}
 	}
 
+	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		const formattedMessages = convertToOpenAiMessages(messages)
 		const systemMessage: OpenAI.Chat.ChatCompletionSystemMessageParam = {

src/api/providers/lmstudio.ts

@@ -4,6 +4,7 @@ import { ApiHandler } from "../"
 import { ApiHandlerOptions, ModelInfo, openAiModelInfoSaneDefaults } from "@shared/api"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"
+import { withRetry } from "../retry"
 
 export class LmStudioHandler implements ApiHandler {
 	private options: ApiHandlerOptions
@@ -17,6 +18,7 @@ export class LmStudioHandler implements ApiHandler {
 		})
 	}
 
+	@withRetry({ retryAllErrors: true })
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
 			{ role: "system", content: systemPrompt },

src/api/providers/qwen.ts

@@ -14,6 +14,7 @@ import {
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"
 import { convertToR1Format } from "../transform/r1-format"
+import { withRetry } from "../retry"
 
 export class QwenHandler implements ApiHandler {
 	private options: ApiHandlerOptions
@@ -48,6 +49,7 @@ export class QwenHandler implements ApiHandler {
 		}
 	}
 
+	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		const model = this.getModel()
 		const isDeepseekReasoner = model.id.includes("deepseek-r1")

src/api/providers/vscode-lm.ts

@@ -7,6 +7,7 @@ import { convertToVsCodeLmMessages } from "@api/transform/vscode-lm-format"
 import { SELECTOR_SEPARATOR, stringifyVsCodeLmModelSelector } from "@shared/vsCodeSelectorUtils"
 import { ApiHandlerOptions, ModelInfo, openAiModelInfoSaneDefaults } from "@shared/api"
 import type { LanguageModelChatSelector as LanguageModelChatSelectorFromTypes } from "./types"
+import { withRetry } from "../retry"
 
 // Cline does not update VSCode type definitions or engine requirements to maintain compatibility.
 // This declaration (as seen in src/integrations/TerminalManager.ts) provides types for the Language Model API in newer versions of VSCode.
@@ -406,6 +407,7 @@ export class VsCodeLmHandler implements ApiHandler, SingleCompletionHandler {
 		return content
 	}
 
+	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		// Ensure clean state before starting a new request
 		this.ensureCleanState()

src/api/providers/xai.ts

@@ -5,6 +5,7 @@ import { ApiHandlerOptions, XAIModelId, ModelInfo, xaiDefaultModelId, xaiModels
 import { convertToOpenAiMessages } from "@api/transform/openai-format"
 import { ApiStream } from "@api/transform/stream"
 import { ChatCompletionReasoningEffort } from "openai/resources/chat/completions"
+import { withRetry } from "../retry"
 
 export class XAIHandler implements ApiHandler {
 	private options: ApiHandlerOptions
@@ -18,6 +19,7 @@ export class XAIHandler implements ApiHandler {
 		})
 	}
 
+	@withRetry()
 	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		const modelId = this.getModel().id
 		// ensure reasoning effort is either "low" or "high" for grok-3-mini