Improve quota- and resource-related 429 error handling, also taking Code Assist customer tiers into consideration (#3609)

packages/cli/src/ui/utils/errorParsing.ts

@@ -4,66 +4,118 @@
  * SPDX-License-Identifier: Apache-2.0
  */
 
-import { AuthType, StructuredError } from '@google/gemini-cli-core';
+import {
+  AuthType,
+  UserTierId,
+  DEFAULT_GEMINI_FLASH_MODEL,
+  DEFAULT_GEMINI_MODEL,
+  isProQuotaExceededError,
+  isGenericQuotaExceededError,
+  isApiError,
+  isStructuredError,
+} from '@google/gemini-cli-core';
 
-const RATE_LIMIT_ERROR_MESSAGE_GOOGLE =
-  '\nPlease wait and try again later. To increase your limits, upgrade to a plan with higher limits, or use /auth to switch to using a paid API key from AI Studio at https://aistudio.google.com/apikey';
+// Free Tier message functions
+const getRateLimitErrorMessageGoogleFree = (
+  fallbackModel: string = DEFAULT_GEMINI_FLASH_MODEL,
+) =>
+  `\nSlow response times detected. Switching to the ${fallbackModel} model for the rest of this session.`;
+
+const getRateLimitErrorMessageGoogleProQuotaFree = (
+  currentModel: string = DEFAULT_GEMINI_MODEL,
+  fallbackModel: string = DEFAULT_GEMINI_FLASH_MODEL,
+) =>
+  `\nYou have reached your daily ${currentModel} quota limit. You will be switched to the ${fallbackModel} model for the rest of this session. To increase your limits, upgrade to a Gemini Code Assist Standard or Enterprise plan with higher limits at https://goo.gle/set-up-gemini-code-assist, or use /auth to switch to using a paid API key from AI Studio at https://aistudio.google.com/apikey`;
+
+const getRateLimitErrorMessageGoogleGenericQuotaFree = () =>
+  `\nYou have reached your daily quota limit. To increase your limits, upgrade to a Gemini Code Assist Standard or Enterprise plan with higher limits at https://goo.gle/set-up-gemini-code-assist, or use /auth to switch to using a paid API key from AI Studio at https://aistudio.google.com/apikey`;
+
+// Legacy/Standard Tier message functions
+const getRateLimitErrorMessageGooglePaid = (
+  fallbackModel: string = DEFAULT_GEMINI_FLASH_MODEL,
+) =>
+  `\nSlow response times detected. Switching to the ${fallbackModel} model for the rest of this session. We appreciate you for choosing Gemini Code Assist and the Gemini CLI.`;
+
+const getRateLimitErrorMessageGoogleProQuotaPaid = (
+  currentModel: string = DEFAULT_GEMINI_MODEL,
+  fallbackModel: string = DEFAULT_GEMINI_FLASH_MODEL,
+) =>
+  `\nYou have reached your daily ${currentModel} quota limit. You will be switched to the ${fallbackModel} model for the rest of this session. We appreciate you for choosing Gemini Code Assist and the Gemini CLI. To continue accessing the ${currentModel} model today, consider using /auth to switch to using a paid API key from AI Studio at https://aistudio.google.com/apikey`;
+
+const getRateLimitErrorMessageGoogleGenericQuotaPaid = (
+  currentModel: string = DEFAULT_GEMINI_MODEL,
+) =>
+  `\nYou have reached your daily quota limit. We appreciate you for choosing Gemini Code Assist and the Gemini CLI. To continue accessing the ${currentModel} model today, consider using /auth to switch to using a paid API key from AI Studio at https://aistudio.google.com/apikey`;
 const RATE_LIMIT_ERROR_MESSAGE_USE_GEMINI =
   '\nPlease wait and try again later. To increase your limits, request a quota increase through AI Studio, or switch to another /auth method';
 const RATE_LIMIT_ERROR_MESSAGE_VERTEX =
   '\nPlease wait and try again later. To increase your limits, request a quota increase through Vertex, or switch to another /auth method';
-const RATE_LIMIT_ERROR_MESSAGE_DEFAULT =
-  'Your request has been rate limited. Please wait and try again later.';
+const getRateLimitErrorMessageDefault = (
+  fallbackModel: string = DEFAULT_GEMINI_FLASH_MODEL,
+) =>
+  `\nSlow response times detected. Switching to the ${fallbackModel} model for the rest of this session.`;
 
-export interface ApiError {
-  error: {
-    code: number;
-    message: string;
-    status: string;
-    details: unknown[];
-  };
-}
-
-function isApiError(error: unknown): error is ApiError {
-  return (
-    typeof error === 'object' &&
-    error !== null &&
-    'error' in error &&
-    typeof (error as ApiError).error === 'object' &&
-    'message' in (error as ApiError).error
-  );
-}
-
-function isStructuredError(error: unknown): error is StructuredError {
-  return (
-    typeof error === 'object' &&
-    error !== null &&
-    'message' in error &&
-    typeof (error as StructuredError).message === 'string'
-  );
-}
-
-function getRateLimitMessage(authType?: AuthType): string {
+function getRateLimitMessage(
+  authType?: AuthType,
+  error?: unknown,
+  userTier?: UserTierId,
+  currentModel?: string,
+  fallbackModel?: string,
+): string {
   switch (authType) {
-    case AuthType.LOGIN_WITH_GOOGLE:
-      return RATE_LIMIT_ERROR_MESSAGE_GOOGLE;
+    case AuthType.LOGIN_WITH_GOOGLE: {
+      // Determine if user is on a paid tier (Legacy or Standard) - default to FREE if not specified
+      const isPaidTier =
+        userTier === UserTierId.LEGACY || userTier === UserTierId.STANDARD;
+
+      if (isProQuotaExceededError(error)) {
+        return isPaidTier
+          ? getRateLimitErrorMessageGoogleProQuotaPaid(
+              currentModel || DEFAULT_GEMINI_MODEL,
+              fallbackModel,
+            )
+          : getRateLimitErrorMessageGoogleProQuotaFree(
+              currentModel || DEFAULT_GEMINI_MODEL,
+              fallbackModel,
+            );
+      } else if (isGenericQuotaExceededError(error)) {
+        return isPaidTier
+          ? getRateLimitErrorMessageGoogleGenericQuotaPaid(
+              currentModel || DEFAULT_GEMINI_MODEL,
+            )
+          : getRateLimitErrorMessageGoogleGenericQuotaFree();
+      } else {
+        return isPaidTier
+          ? getRateLimitErrorMessageGooglePaid(fallbackModel)
+          : getRateLimitErrorMessageGoogleFree(fallbackModel);
+      }
+    }
     case AuthType.USE_GEMINI:
       return RATE_LIMIT_ERROR_MESSAGE_USE_GEMINI;
     case AuthType.USE_VERTEX_AI:
       return RATE_LIMIT_ERROR_MESSAGE_VERTEX;
     default:
-      return RATE_LIMIT_ERROR_MESSAGE_DEFAULT;
+      return getRateLimitErrorMessageDefault(fallbackModel);
   }
 }
 
 export function parseAndFormatApiError(
   error: unknown,
   authType?: AuthType,
+  userTier?: UserTierId,
+  currentModel?: string,
+  fallbackModel?: string,
 ): string {
   if (isStructuredError(error)) {
     let text = `[API Error: ${error.message}]`;
     if (error.status === 429) {
-      text += getRateLimitMessage(authType);
+      text += getRateLimitMessage(
+        authType,
+        error,
+        userTier,
+        currentModel,
+        fallbackModel,
+      );
     }
     return text;
   }
@@ -92,7 +144,13 @@ export function parseAndFormatApiError(
         }
         let text = `[API Error: ${finalMessage} (Status: ${parsedError.error.status})]`;
         if (parsedError.error.code === 429) {
-          text += getRateLimitMessage(authType);
+          text += getRateLimitMessage(
+            authType,
+            parsedError,
+            userTier,
+            currentModel,
+            fallbackModel,
+          );
         }
         return text;
       }