abbyswag/qwen-code
1
0
Fork 0
mirror of https://github.com/QwenLM/qwen-code.git synced 2025-12-20 16:57:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

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

Browse Source
This commit is contained in:
Bryan Morgan
2025-07-09 10:18:15 -04:00
committed by GitHub
parent 8f2da86aa5
commit b0cce95286
14 changed files with 611 additions and 63 deletions
Download Patch File Download Diff File Expand all files Collapse all files

309
packages/cli/src/ui/utils/errorParsing.test.ts
View File

@@ -6,10 +6,16 @@
import { describe, it, expect } from 'vitest';
import { parseAndFormatApiError } from './errorParsing.js';
import { AuthType, StructuredError } from '@google/gemini-cli-core';
import {
AuthType,
UserTierId,
DEFAULT_GEMINI_FLASH_MODEL,
isProQuotaExceededError,
} from '@google/gemini-cli-core';
describe('parseAndFormatApiError', () => {
const enterpriseMessage = 'upgrade to a plan with higher limits';
const _enterpriseMessage =
'upgrade to a Gemini Code Assist Standard or Enterprise plan with higher limits';
const vertexMessage = 'request a quota increase through Vertex';
const geminiMessage = 'request a quota increase through AI Studio';
@@ -24,9 +30,17 @@ describe('parseAndFormatApiError', () => {
it('should format a 429 API error with the default message', () => {
const errorMessage =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Rate limit exceeded","status":"RESOURCE_EXHAUSTED"}}';
const result = parseAndFormatApiError(errorMessage);
const result = parseAndFormatApiError(
errorMessage,
undefined,
undefined,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain('[API Error: Rate limit exceeded');
expect(result).toContain('Your request has been rate limited');
expect(result).toContain(
'Slow response times detected. Switching to the gemini-2.5-flash model',
);
});
it('should format a 429 API error with the personal message', () => {
@@ -35,9 +49,14 @@ describe('parseAndFormatApiError', () => {
const result = parseAndFormatApiError(
errorMessage,
AuthType.LOGIN_WITH_GOOGLE,
undefined,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain('[API Error: Rate limit exceeded');
expect(result).toContain(enterpriseMessage);
expect(result).toContain(
'Slow response times detected. Switching to the gemini-2.5-flash model',
);
});
it('should format a 429 API error with the vertex message', () => {
@@ -116,4 +135,284 @@ describe('parseAndFormatApiError', () => {
const expected = '[API Error: An unknown error occurred.]';
expect(parseAndFormatApiError(error)).toBe(expected);
});
it('should format a 429 API error with Pro quota exceeded message for Google auth (Free tier)', () => {
const errorMessage =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'Gemini 2.5 Pro Requests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const result = parseAndFormatApiError(
errorMessage,
AuthType.LOGIN_WITH_GOOGLE,
undefined,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain(
"[API Error: Quota exceeded for quota metric 'Gemini 2.5 Pro Requests'",
);
expect(result).toContain(
'You have reached your daily gemini-2.5-pro quota limit',
);
expect(result).toContain(
'upgrade to a Gemini Code Assist Standard or Enterprise plan',
);
});
it('should format a regular 429 API error with standard message for Google auth', () => {
const errorMessage =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Rate limit exceeded","status":"RESOURCE_EXHAUSTED"}}';
const result = parseAndFormatApiError(
errorMessage,
AuthType.LOGIN_WITH_GOOGLE,
undefined,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain('[API Error: Rate limit exceeded');
expect(result).toContain(
'Slow response times detected. Switching to the gemini-2.5-flash model',
);
expect(result).not.toContain(
'You have reached your daily gemini-2.5-pro quota limit',
);
});
it('should format a 429 API error with generic quota exceeded message for Google auth', () => {
const errorMessage =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'GenerationRequests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const result = parseAndFormatApiError(
errorMessage,
AuthType.LOGIN_WITH_GOOGLE,
undefined,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain(
"[API Error: Quota exceeded for quota metric 'GenerationRequests'",
);
expect(result).toContain('You have reached your daily quota limit');
expect(result).not.toContain(
'You have reached your daily Gemini 2.5 Pro quota limit',
);
});
it('should prioritize Pro quota message over generic quota message for Google auth', () => {
const errorMessage =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'Gemini 2.5 Pro Requests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const result = parseAndFormatApiError(
errorMessage,
AuthType.LOGIN_WITH_GOOGLE,
undefined,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain(
"[API Error: Quota exceeded for quota metric 'Gemini 2.5 Pro Requests'",
);
expect(result).toContain(
'You have reached your daily gemini-2.5-pro quota limit',
);
expect(result).not.toContain('You have reached your daily quota limit');
});
it('should format a 429 API error with Pro quota exceeded message for Google auth (Standard tier)', () => {
const errorMessage =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'Gemini 2.5 Pro Requests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const result = parseAndFormatApiError(
errorMessage,
AuthType.LOGIN_WITH_GOOGLE,
UserTierId.STANDARD,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain(
"[API Error: Quota exceeded for quota metric 'Gemini 2.5 Pro Requests'",
);
expect(result).toContain(
'You have reached your daily gemini-2.5-pro quota limit',
);
expect(result).toContain(
'We appreciate you for choosing Gemini Code Assist and the Gemini CLI',
);
expect(result).not.toContain(
'upgrade to a Gemini Code Assist Standard or Enterprise plan',
);
});
it('should format a 429 API error with Pro quota exceeded message for Google auth (Legacy tier)', () => {
const errorMessage =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'Gemini 2.5 Pro Requests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const result = parseAndFormatApiError(
errorMessage,
AuthType.LOGIN_WITH_GOOGLE,
UserTierId.LEGACY,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain(
"[API Error: Quota exceeded for quota metric 'Gemini 2.5 Pro Requests'",
);
expect(result).toContain(
'You have reached your daily gemini-2.5-pro quota limit',
);
expect(result).toContain(
'We appreciate you for choosing Gemini Code Assist and the Gemini CLI',
);
expect(result).not.toContain(
'upgrade to a Gemini Code Assist Standard or Enterprise plan',
);
});
it('should handle different Gemini version strings in Pro quota exceeded errors', () => {
const errorMessage15 =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'Gemini 1.5 Pro Requests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const errorMessagePreview =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'Gemini 2.5-preview Pro Requests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const errorMessageBeta =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'Gemini beta-3.0 Pro Requests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const errorMessageExperimental =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'Gemini experimental-v2 Pro Requests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const result15 = parseAndFormatApiError(
errorMessage15,
AuthType.LOGIN_WITH_GOOGLE,
undefined,
'gemini-1.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
const resultPreview = parseAndFormatApiError(
errorMessagePreview,
AuthType.LOGIN_WITH_GOOGLE,
undefined,
'gemini-2.5-preview-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
const resultBeta = parseAndFormatApiError(
errorMessageBeta,
AuthType.LOGIN_WITH_GOOGLE,
undefined,
'gemini-beta-3.0-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
const resultExperimental = parseAndFormatApiError(
errorMessageExperimental,
AuthType.LOGIN_WITH_GOOGLE,
undefined,
'gemini-experimental-v2-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result15).toContain(
'You have reached your daily gemini-1.5-pro quota limit',
);
expect(resultPreview).toContain(
'You have reached your daily gemini-2.5-preview-pro quota limit',
);
expect(resultBeta).toContain(
'You have reached your daily gemini-beta-3.0-pro quota limit',
);
expect(resultExperimental).toContain(
'You have reached your daily gemini-experimental-v2-pro quota limit',
);
expect(result15).toContain(
'upgrade to a Gemini Code Assist Standard or Enterprise plan',
);
expect(resultPreview).toContain(
'upgrade to a Gemini Code Assist Standard or Enterprise plan',
);
expect(resultBeta).toContain(
'upgrade to a Gemini Code Assist Standard or Enterprise plan',
);
expect(resultExperimental).toContain(
'upgrade to a Gemini Code Assist Standard or Enterprise plan',
);
});
it('should not match non-Pro models with similar version strings', () => {
// Test that Flash models with similar version strings don't match
expect(
isProQuotaExceededError(
"Quota exceeded for quota metric 'Gemini 2.5 Flash Requests' and limit",
),
).toBe(false);
expect(
isProQuotaExceededError(
"Quota exceeded for quota metric 'Gemini 2.5-preview Flash Requests' and limit",
),
).toBe(false);
expect(
isProQuotaExceededError(
"Quota exceeded for quota metric 'Gemini beta-3.0 Flash Requests' and limit",
),
).toBe(false);
expect(
isProQuotaExceededError(
"Quota exceeded for quota metric 'Gemini experimental-v2 Flash Requests' and limit",
),
).toBe(false);
// Test other model types
expect(
isProQuotaExceededError(
"Quota exceeded for quota metric 'Gemini 2.5 Ultra Requests' and limit",
),
).toBe(false);
expect(
isProQuotaExceededError(
"Quota exceeded for quota metric 'Gemini 2.5 Standard Requests' and limit",
),
).toBe(false);
// Test generic quota messages
expect(
isProQuotaExceededError(
"Quota exceeded for quota metric 'GenerationRequests' and limit",
),
).toBe(false);
expect(
isProQuotaExceededError(
"Quota exceeded for quota metric 'EmbeddingRequests' and limit",
),
).toBe(false);
});
it('should format a generic quota exceeded message for Google auth (Standard tier)', () => {
const errorMessage =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Quota exceeded for quota metric \'GenerationRequests\' and limit \'RequestsPerDay\' of service \'generativelanguage.googleapis.com\' for consumer \'project_number:123456789\'.","status":"RESOURCE_EXHAUSTED"}}';
const result = parseAndFormatApiError(
errorMessage,
AuthType.LOGIN_WITH_GOOGLE,
UserTierId.STANDARD,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain(
"[API Error: Quota exceeded for quota metric 'GenerationRequests'",
);
expect(result).toContain('You have reached your daily quota limit');
expect(result).toContain(
'We appreciate you for choosing Gemini Code Assist and the Gemini CLI',
);
expect(result).not.toContain(
'upgrade to a Gemini Code Assist Standard or Enterprise plan',
);
});
it('should format a regular 429 API error with standard message for Google auth (Standard tier)', () => {
const errorMessage =
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Rate limit exceeded","status":"RESOURCE_EXHAUSTED"}}';
const result = parseAndFormatApiError(
errorMessage,
AuthType.LOGIN_WITH_GOOGLE,
UserTierId.STANDARD,
'gemini-2.5-pro',
DEFAULT_GEMINI_FLASH_MODEL,
);
expect(result).toContain('[API Error: Rate limit exceeded');
expect(result).toContain(
'We appreciate you for choosing Gemini Code Assist and the Gemini CLI',
);
expect(result).not.toContain(
'upgrade to a Gemini Code Assist Standard or Enterprise plan',
);
});
});

136
packages/cli/src/ui/utils/errorParsing.ts
View File

@@ -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;
}