🛡️ feat: Rate Limiting for Conversation Forking (danny-avila#8269)

* chore: Improve error logging for fetching conversations, and use new TS packages for utils

* feat: Implement fork limiters for conversation forking requests

* chore: error message for conversation index deletion to clarify syncing behavior

* feat: Enhance error handling for forking with rate limit message

Author: danny-avila

api/server/middleware/limiters/forkLimiters.js

@@ -0,0 +1,88 @@
+const rateLimit = require('express-rate-limit');
+const { isEnabled } = require('@librechat/api');
+const { RedisStore } = require('rate-limit-redis');
+const { logger } = require('@librechat/data-schemas');
+const { ViolationTypes } = require('librechat-data-provider');
+const ioredisClient = require('~/cache/ioredisClient');
+const logViolation = require('~/cache/logViolation');
+
+const getEnvironmentVariables = () => {
+  const FORK_IP_MAX = parseInt(process.env.FORK_IP_MAX) || 30;
+  const FORK_IP_WINDOW = parseInt(process.env.FORK_IP_WINDOW) || 1;
+  const FORK_USER_MAX = parseInt(process.env.FORK_USER_MAX) || 7;
+  const FORK_USER_WINDOW = parseInt(process.env.FORK_USER_WINDOW) || 1;
+
+  const forkIpWindowMs = FORK_IP_WINDOW * 60 * 1000;
+  const forkIpMax = FORK_IP_MAX;
+  const forkIpWindowInMinutes = forkIpWindowMs / 60000;
+
+  const forkUserWindowMs = FORK_USER_WINDOW * 60 * 1000;
+  const forkUserMax = FORK_USER_MAX;
+  const forkUserWindowInMinutes = forkUserWindowMs / 60000;
+
+  return {
+    forkIpWindowMs,
+    forkIpMax,
+    forkIpWindowInMinutes,
+    forkUserWindowMs,
+    forkUserMax,
+    forkUserWindowInMinutes,
+  };
+};
+
+const createForkHandler = (ip = true) => {
+  const { forkIpMax, forkIpWindowInMinutes, forkUserMax, forkUserWindowInMinutes } =
+    getEnvironmentVariables();
+
+  return async (req, res) => {
+    const type = ViolationTypes.FILE_UPLOAD_LIMIT;
+    const errorMessage = {
+      type,
+      max: ip ? forkIpMax : forkUserMax,
+      limiter: ip ? 'ip' : 'user',
+      windowInMinutes: ip ? forkIpWindowInMinutes : forkUserWindowInMinutes,
+    };
+
+    await logViolation(req, res, type, errorMessage);
+    res.status(429).json({ message: 'Too many conversation fork requests. Try again later' });
+  };
+};
+
+const createForkLimiters = () => {
+  const { forkIpWindowMs, forkIpMax, forkUserWindowMs, forkUserMax } = getEnvironmentVariables();
+
+  const ipLimiterOptions = {
+    windowMs: forkIpWindowMs,
+    max: forkIpMax,
+    handler: createForkHandler(),
+  };
+  const userLimiterOptions = {
+    windowMs: forkUserWindowMs,
+    max: forkUserMax,
+    handler: createForkHandler(false),
+    keyGenerator: function (req) {
+      return req.user?.id;
+    },
+  };
+
+  if (isEnabled(process.env.USE_REDIS) && ioredisClient) {
+    logger.debug('Using Redis for fork rate limiters.');
+    const sendCommand = (...args) => ioredisClient.call(...args);
+    const ipStore = new RedisStore({
+      sendCommand,
+      prefix: 'fork_ip_limiter:',
+    });
+    const userStore = new RedisStore({
+      sendCommand,
+      prefix: 'fork_user_limiter:',
+    });
+    ipLimiterOptions.store = ipStore;
+    userLimiterOptions.store = userStore;
+  }
+
+  const forkIpLimiter = rateLimit(ipLimiterOptions);
+  const forkUserLimiter = rateLimit(userLimiterOptions);
+  return { forkIpLimiter, forkUserLimiter };
+};
+
+module.exports = { createForkLimiters };

api/server/middleware/limiters/importLimiters.js

@@ -1,10 +1,10 @@
 const rateLimit = require('express-rate-limit');
+const { isEnabled } = require('@librechat/api');
 const { RedisStore } = require('rate-limit-redis');
+const { logger } = require('@librechat/data-schemas');
 const { ViolationTypes } = require('librechat-data-provider');
 const ioredisClient = require('~/cache/ioredisClient');
 const logViolation = require('~/cache/logViolation');
-const { isEnabled } = require('~/server/utils');
-const { logger } = require('~/config');
 
 const getEnvironmentVariables = () => {
   const IMPORT_IP_MAX = parseInt(process.env.IMPORT_IP_MAX) || 100;

api/server/middleware/limiters/index.js

@@ -4,6 +4,7 @@ const createSTTLimiters = require('./sttLimiters');
 const loginLimiter = require('./loginLimiter');
 const importLimiters = require('./importLimiters');
 const uploadLimiters = require('./uploadLimiters');
+const forkLimiters = require('./forkLimiters');
 const registerLimiter = require('./registerLimiter');
 const toolCallLimiter = require('./toolCallLimiter');
 const messageLimiters = require('./messageLimiters');
@@ -14,6 +15,7 @@ module.exports = {
   ...uploadLimiters,
   ...importLimiters,
   ...messageLimiters,
+  ...forkLimiters,
   loginLimiter,
   registerLimiter,
   toolCallLimiter,

api/server/routes/convos.js

@@ -1,16 +1,17 @@
 const multer = require('multer');
 const express = require('express');
+const { sleep } = require('@librechat/agents');
+const { isEnabled } = require('@librechat/api');
+const { logger } = require('@librechat/data-schemas');
 const { CacheKeys, EModelEndpoint } = require('librechat-data-provider');
 const { getConvosByCursor, deleteConvos, getConvo, saveConvo } = require('~/models/Conversation');
 const { forkConversation, duplicateConversation } = require('~/server/utils/import/fork');
+const { createImportLimiters, createForkLimiters } = require('~/server/middleware');
 const { storage, importFileFilter } = require('~/server/routes/files/multer');
 const requireJwtAuth = require('~/server/middleware/requireJwtAuth');
 const { importConversations } = require('~/server/utils/import');
-const { createImportLimiters } = require('~/server/middleware');
 const { deleteToolCalls } = require('~/models/ToolCall');
-const { isEnabled, sleep } = require('~/server/utils');
 const getLogStores = require('~/cache/getLogStores');
-const { logger } = require('~/config');
 
 const assistantClients = {
   [EModelEndpoint.azureAssistants]: require('~/server/services/Endpoints/azureAssistants'),
@@ -43,6 +44,7 @@ router.get('/', async (req, res) => {
     });
     res.status(200).json(result);
   } catch (error) {
+    logger.error('Error fetching conversations', error);
     res.status(500).json({ error: 'Error fetching conversations' });
   }
 });
@@ -156,6 +158,7 @@ router.post('/update', async (req, res) => {
 });
 
 const { importIpLimiter, importUserLimiter } = createImportLimiters();
+const { forkIpLimiter, forkUserLimiter } = createForkLimiters();
 const upload = multer({ storage: storage, fileFilter: importFileFilter });
 
 /**
@@ -189,7 +192,7 @@ router.post(
  * @param {express.Response<TForkConvoResponse>} res - Express response object.
  * @returns {Promise<void>} - The response after forking the conversation.
  */
-router.post('/fork', async (req, res) => {
+router.post('/fork', forkIpLimiter, forkUserLimiter, async (req, res) => {
   try {
     /** @type {TForkConvoRequest} */
     const { conversationId, messageId, option, splitAtTarget, latestMessageId } = req.body;

client/src/components/Chat/Messages/Fork.tsx

@@ -233,9 +233,17 @@ export default function Fork({
         status: 'info',
       });
     },
-    onError: () => {
+    onError: (error) => {
+      /** Rate limit error (429 status code) */
+      const isRateLimitError =
+        (error as any)?.response?.status === 429 ||
+        (error as any)?.status === 429 ||
+        (error as any)?.statusCode === 429;
+
       showToast({
-        message: localize('com_ui_fork_error'),
+        message: isRateLimitError
+          ? localize('com_ui_fork_error_rate_limit')
+          : localize('com_ui_fork_error'),
         status: 'error',
       });
     },

client/src/locales/en/translation.json

@@ -775,6 +775,7 @@
   "com_ui_fork_change_default": "Default fork option",
   "com_ui_fork_default": "Use default fork option",
   "com_ui_fork_error": "There was an error forking the conversation",
+  "com_ui_fork_error_rate_limit": "Too many fork requests. Please try again later",
   "com_ui_fork_from_message": "Select a fork option",
   "com_ui_fork_info_1": "Use this setting to fork messages with the desired behavior.",
   "com_ui_fork_info_2": "\"Forking\" refers to creating a new conversation that start/end from specific messages in the current conversation, creating a copy according to the options selected.",

packages/data-schemas/src/models/plugins/mongoMeili.ts

@@ -716,7 +716,7 @@ export default function mongoMeili(schema: Schema, options: MongoMeiliOptions):
     } catch (error) {
       if (meiliEnabled) {
         logger.error(
-          '[MeiliMongooseModel.deleteMany] There was an issue deleting conversation indexes upon deletion. Next startup may be slow due to syncing.',
+          '[MeiliMongooseModel.deleteMany] There was an issue deleting conversation indexes upon deletion. Next startup may trigger syncing.',
           error,
         );
       }