mirror of
https://github.com/google-gemini/gemini-cli.git
synced 2026-04-27 05:24:34 -07:00
fix: handle request retries and model fallback correctly (#11624)
This commit is contained in:
@@ -6,9 +6,7 @@
|
||||
|
||||
import { describe, it, expect } from 'vitest';
|
||||
import { parseAndFormatApiError } from './errorParsing.js';
|
||||
import { isProQuotaExceededError } from './quotaErrorDetection.js';
|
||||
import { DEFAULT_GEMINI_FLASH_MODEL } from '../config/models.js';
|
||||
import { UserTierId } from '../code_assist/types.js';
|
||||
import { AuthType } from '../core/contentGenerator.js';
|
||||
import type { StructuredError } from '../core/turn.js';
|
||||
|
||||
@@ -40,22 +38,6 @@ describe('parseAndFormatApiError', () => {
|
||||
);
|
||||
});
|
||||
|
||||
it('should format a 429 API error with the personal message', () => {
|
||||
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(
|
||||
'Possible quota limitations in place or slow response times detected. Switching to the gemini-2.5-flash model',
|
||||
);
|
||||
});
|
||||
|
||||
it('should format a 429 API error with the vertex message', () => {
|
||||
const errorMessage =
|
||||
'got status: 429 Too Many Requests. {"error":{"code":429,"message":"Rate limit exceeded","status":"RESOURCE_EXHAUSTED"}}';
|
||||
@@ -132,230 +114,4 @@ 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 get higher limits');
|
||||
});
|
||||
|
||||
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(
|
||||
'Possible quota limitations in place or 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 get higher limits');
|
||||
});
|
||||
|
||||
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 get higher limits');
|
||||
});
|
||||
|
||||
it('should handle different Gemini 2.5 version strings in Pro quota exceeded errors', () => {
|
||||
const errorMessage25 =
|
||||
'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 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 result25 = parseAndFormatApiError(
|
||||
errorMessage25,
|
||||
AuthType.LOGIN_WITH_GOOGLE,
|
||||
undefined,
|
||||
'gemini-2.5-pro',
|
||||
DEFAULT_GEMINI_FLASH_MODEL,
|
||||
);
|
||||
const resultPreview = parseAndFormatApiError(
|
||||
errorMessagePreview,
|
||||
AuthType.LOGIN_WITH_GOOGLE,
|
||||
undefined,
|
||||
'gemini-2.5-preview-pro',
|
||||
DEFAULT_GEMINI_FLASH_MODEL,
|
||||
);
|
||||
|
||||
expect(result25).toContain(
|
||||
'You have reached your daily gemini-2.5-pro quota limit',
|
||||
);
|
||||
expect(resultPreview).toContain(
|
||||
'You have reached your daily gemini-2.5-preview-pro quota limit',
|
||||
);
|
||||
expect(result25).toContain('upgrade to get higher limits');
|
||||
expect(resultPreview).toContain('upgrade to get higher limits');
|
||||
});
|
||||
|
||||
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);
|
||||
|
||||
// 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 get higher limits');
|
||||
});
|
||||
|
||||
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 get higher limits');
|
||||
});
|
||||
});
|
||||
|
||||
Reference in New Issue
Block a user