feat: implement dynamic retry timing for Google Gemini API

- Enhanced error handling in GeminiHandler to preserve original error structure
- Added logic to distinguish between rate limits and quota exhaustion
- Preserved retry delay information from Gemini API responses
- Added new localized error messages for better user feedback
- Added comprehensive tests for the new retry logic

Fixes #6680

Author: roomote

src/api/providers/__tests__/gemini.spec.ts

@@ -102,6 +102,67 @@ describe("GeminiHandler", () => {
 				}
 			}).rejects.toThrow()
 		})
+
+		it("should handle rate limit errors with retry information", async () => {
+			const mockError: any = new Error("Rate limit exceeded")
+			mockError.status = 429
+			mockError.errorDetails = [
+				{
+					"@type": "type.googleapis.com/google.rpc.RetryInfo",
+					retryDelay: "40s",
+				},
+			]
+			;(handler["client"].models.generateContentStream as any).mockRejectedValue(mockError)
+
+			const stream = handler.createMessage(systemPrompt, mockMessages)
+
+			await expect(async () => {
+				for await (const _chunk of stream) {
+					// Should throw before yielding any chunks
+				}
+			}).rejects.toThrow(t("common:errors.gemini.rate_limit"))
+		})
+
+		it("should handle quota exhaustion errors", async () => {
+			const mockError: any = new Error(
+				"You exceeded your current quota, please check your plan and billing details",
+			)
+			mockError.status = 429
+			mockError.message = "You exceeded your current quota, please check your plan and billing details"
+			;(handler["client"].models.generateContentStream as any).mockRejectedValue(mockError)
+
+			const stream = handler.createMessage(systemPrompt, mockMessages)
+
+			await expect(async () => {
+				for await (const _chunk of stream) {
+					// Should throw before yielding any chunks
+				}
+			}).rejects.toThrow(t("common:errors.gemini.quota_exhausted"))
+		})
+
+		it("should preserve error details for retry logic", async () => {
+			const mockError: any = new Error("Rate limit exceeded")
+			mockError.status = 429
+			mockError.errorDetails = [
+				{
+					"@type": "type.googleapis.com/google.rpc.RetryInfo",
+					retryDelay: "60s",
+				},
+			]
+			;(handler["client"].models.generateContentStream as any).mockRejectedValue(mockError)
+
+			const stream = handler.createMessage(systemPrompt, mockMessages)
+
+			try {
+				for await (const _chunk of stream) {
+					// Should throw before yielding any chunks
+				}
+			} catch (error: any) {
+				expect(error.status).toBe(429)
+				expect(error.errorDetails).toBeDefined()
+				expect(error.errorDetails[0].retryDelay).toBe("60s")
+			}
+		})
 	})
 
 	describe("completePrompt", () => {
@@ -134,6 +195,33 @@ describe("GeminiHandler", () => {
 			)
 		})
 
+		it("should handle rate limit errors in completePrompt", async () => {
+			const mockError: any = new Error("Rate limit exceeded")
+			mockError.status = 429
+			mockError.errorDetails = [
+				{
+					"@type": "type.googleapis.com/google.rpc.RetryInfo",
+					retryDelay: "30s",
+				},
+			]
+			;(handler["client"].models.generateContent as any).mockRejectedValue(mockError)
+
+			await expect(handler.completePrompt("Test prompt")).rejects.toThrow(t("common:errors.gemini.rate_limit"))
+		})
+
+		it("should handle quota exhaustion errors in completePrompt", async () => {
+			const mockError: any = new Error(
+				"You exceeded your current quota, please check your plan and billing details",
+			)
+			mockError.status = 429
+			mockError.message = "You exceeded your current quota, please check your plan and billing details"
+			;(handler["client"].models.generateContent as any).mockRejectedValue(mockError)
+
+			await expect(handler.completePrompt("Test prompt")).rejects.toThrow(
+				t("common:errors.gemini.quota_exhausted"),
+			)
+		})
+
 		it("should handle empty response", async () => {
 			// Mock the response with empty text
 			;(handler["client"].models.generateContent as any).mockResolvedValue({

src/api/providers/gemini.ts

@@ -153,7 +153,30 @@ export class GeminiHandler extends BaseProvider implements SingleCompletionHandl
 					totalCost: this.calculateCost({ info, inputTokens, outputTokens, cacheReadTokens }),
 				}
 			}
-		} catch (error) {
+		} catch (error: any) {
+			// Preserve the original error structure for retry logic
+			if (error.status === 429) {
+				// Check if this is a rate limit or quota exhaustion
+				const errorMessage = error.message || ""
+				const isQuotaExhausted = errorMessage.includes("quota") || errorMessage.includes("billing")
+
+				// Create an enhanced error that preserves the original structure
+				const enhancedError = new Error(
+					isQuotaExhausted
+						? t("common:errors.gemini.quota_exhausted", { error: errorMessage })
+						: t("common:errors.gemini.rate_limit", { error: errorMessage }),
+				)
+
+				// Preserve the original error properties for retry logic
+				Object.assign(enhancedError, {
+					status: error.status,
+					errorDetails: error.errorDetails,
+					message: error.message,
+				})
+
+				throw enhancedError
+			}
+
 			if (error instanceof Error) {
 				throw new Error(t("common:errors.gemini.generate_stream", { error: error.message }))
 			}
@@ -235,7 +258,30 @@ export class GeminiHandler extends BaseProvider implements SingleCompletionHandl
 			}
 
 			return text
-		} catch (error) {
+		} catch (error: any) {
+			// Preserve the original error structure for retry logic
+			if (error.status === 429) {
+				// Check if this is a rate limit or quota exhaustion
+				const errorMessage = error.message || ""
+				const isQuotaExhausted = errorMessage.includes("quota") || errorMessage.includes("billing")
+
+				// Create an enhanced error that preserves the original structure
+				const enhancedError = new Error(
+					isQuotaExhausted
+						? t("common:errors.gemini.quota_exhausted", { error: errorMessage })
+						: t("common:errors.gemini.rate_limit", { error: errorMessage }),
+				)
+
+				// Preserve the original error properties for retry logic
+				Object.assign(enhancedError, {
+					status: error.status,
+					errorDetails: error.errorDetails,
+					message: error.message,
+				})
+
+				throw enhancedError
+			}
+
 			if (error instanceof Error) {
 				throw new Error(t("common:errors.gemini.generate_complete_prompt", { error: error.message }))
 			}

src/i18n/locales/en/common.json

@@ -89,7 +89,9 @@
 		"gemini": {
 			"generate_stream": "Gemini generate context stream error: {{error}}",
 			"generate_complete_prompt": "Gemini completion error: {{error}}",
-			"sources": "Sources:"
+			"sources": "Sources:",
+			"rate_limit": "Gemini API rate limit exceeded. The system will automatically retry with the appropriate delay.",
+			"quota_exhausted": "Gemini API quota exhausted: {{error}}. Please check your plan and billing details at https://ai.google.dev/gemini-api/docs/rate-limits"
 		},
 		"cerebras": {
 			"authenticationFailed": "Cerebras API authentication failed. Please check your API key is valid and not expired.",