improvement(rooms): redis client closed should fail fast

apps/sim/app/workspace/providers/socket-provider.tsx

@@ -14,6 +14,7 @@ import { createLogger } from '@sim/logger'
 import { useParams } from 'next/navigation'
 import { io, type Socket } from 'socket.io-client'
 import { getEnv } from '@/lib/core/config/env'
+import { useOperationQueueStore } from '@/stores/operation-queue/store'
 
 const logger = createLogger('SocketContext')
 
@@ -138,6 +139,7 @@ export function SocketProvider({ children, user }: SocketProviderProps) {
   const [authFailed, setAuthFailed] = useState(false)
   const initializedRef = useRef(false)
   const socketRef = useRef<Socket | null>(null)
+  const triggerOfflineMode = useOperationQueueStore((state) => state.triggerOfflineMode)
 
   const params = useParams()
   const urlWorkflowId = params?.workflowId as string | undefined
@@ -341,9 +343,12 @@ export function SocketProvider({ children, user }: SocketProviderProps) {
           })
         })
 
-        socketInstance.on('join-workflow-error', ({ error }) => {
+        socketInstance.on('join-workflow-error', ({ error, code }) => {
           isRejoiningRef.current = false
-          logger.error('Failed to join workflow:', error)
+          logger.error('Failed to join workflow:', { error, code })
+          if (code === 'ROOM_MANAGER_UNAVAILABLE') {
+            triggerOfflineMode()
+          }
         })
 
         socketInstance.on('workflow-operation', (data) => {

apps/sim/socket/handlers/operations.ts

@@ -12,15 +12,49 @@ import {
 import { persistWorkflowOperation } from '@/socket/database/operations'
 import type { AuthenticatedSocket } from '@/socket/middleware/auth'
 import { checkRolePermission } from '@/socket/middleware/permissions'
-import type { IRoomManager } from '@/socket/rooms'
+import type { IRoomManager, UserSession } from '@/socket/rooms'
 import { WorkflowOperationSchema } from '@/socket/validation/schemas'
 
 const logger = createLogger('OperationsHandlers')
 
 export function setupOperationsHandlers(socket: AuthenticatedSocket, roomManager: IRoomManager) {
   socket.on('workflow-operation', async (data) => {
-    const workflowId = await roomManager.getWorkflowIdForSocket(socket.id)
-    const session = await roomManager.getUserSession(socket.id)
+    if (!roomManager.isReady()) {
+      socket.emit('operation-forbidden', {
+        type: 'ROOM_MANAGER_UNAVAILABLE',
+        message: 'Realtime unavailable',
+      })
+      if (data?.operationId) {
+        socket.emit('operation-failed', {
+          operationId: data.operationId,
+          error: 'Realtime unavailable',
+          retryable: true,
+        })
+      }
+      return
+    }
+
+    let workflowId: string | null = null
+    let session: UserSession | null = null
+
+    try {
+      workflowId = await roomManager.getWorkflowIdForSocket(socket.id)
+      session = await roomManager.getUserSession(socket.id)
+    } catch (error) {
+      logger.error('Error loading session for workflow operation:', error)
+      socket.emit('operation-forbidden', {
+        type: 'ROOM_MANAGER_UNAVAILABLE',
+        message: 'Realtime unavailable',
+      })
+      if (data?.operationId) {
+        socket.emit('operation-failed', {
+          operationId: data.operationId,
+          error: 'Realtime unavailable',
+          retryable: true,
+        })
+      }
+      return
+    }
 
     if (!workflowId || !session) {
       socket.emit('operation-forbidden', {

apps/sim/socket/handlers/subblocks.ts

@@ -48,6 +48,21 @@ export function setupSubblocksHandlers(socket: AuthenticatedSocket, roomManager:
       operationId,
     } = data
 
+    if (!roomManager.isReady()) {
+      socket.emit('operation-forbidden', {
+        type: 'ROOM_MANAGER_UNAVAILABLE',
+        message: 'Realtime unavailable',
+      })
+      if (operationId) {
+        socket.emit('operation-failed', {
+          operationId,
+          error: 'Realtime unavailable',
+          retryable: true,
+        })
+      }
+      return
+    }
+
     try {
       const sessionWorkflowId = await roomManager.getWorkflowIdForSocket(socket.id)
       const session = await roomManager.getUserSession(socket.id)

apps/sim/socket/handlers/variables.ts

@@ -37,6 +37,21 @@ export function setupVariablesHandlers(socket: AuthenticatedSocket, roomManager:
   socket.on('variable-update', async (data) => {
     const { workflowId: payloadWorkflowId, variableId, field, value, timestamp, operationId } = data
 
+    if (!roomManager.isReady()) {
+      socket.emit('operation-forbidden', {
+        type: 'ROOM_MANAGER_UNAVAILABLE',
+        message: 'Realtime unavailable',
+      })
+      if (operationId) {
+        socket.emit('operation-failed', {
+          operationId,
+          error: 'Realtime unavailable',
+          retryable: true,
+        })
+      }
+      return
+    }
+
     try {
       const sessionWorkflowId = await roomManager.getWorkflowIdForSocket(socket.id)
       const session = await roomManager.getUserSession(socket.id)

apps/sim/socket/handlers/workflow.ts

@@ -20,6 +20,15 @@ export function setupWorkflowHandlers(socket: AuthenticatedSocket, roomManager:
         return
       }
 
+      if (!roomManager.isReady()) {
+        logger.warn(`Join workflow rejected: Room manager unavailable`)
+        socket.emit('join-workflow-error', {
+          error: 'Realtime unavailable',
+          code: 'ROOM_MANAGER_UNAVAILABLE',
+        })
+        return
+      }
+
       logger.info(`Join workflow request from ${userId} (${userName}) for workflow ${workflowId}`)
 
       // Verify workflow access
@@ -128,12 +137,19 @@ export function setupWorkflowHandlers(socket: AuthenticatedSocket, roomManager:
       // Undo socket.join and room manager entry if any operation failed
       socket.leave(workflowId)
       await roomManager.removeUserFromRoom(socket.id)
-      socket.emit('join-workflow-error', { error: 'Failed to join workflow' })
+      socket.emit('join-workflow-error', {
+        error: roomManager.isReady() ? 'Failed to join workflow' : 'Realtime unavailable',
+        code: roomManager.isReady() ? undefined : 'ROOM_MANAGER_UNAVAILABLE',
+      })
     }
   })
 
   socket.on('leave-workflow', async () => {
     try {
+      if (!roomManager.isReady()) {
+        return
+      }
+
       const workflowId = await roomManager.getWorkflowIdForSocket(socket.id)
       const session = await roomManager.getUserSession(socket.id)
 

apps/sim/socket/rooms/memory-manager.ts

@@ -26,6 +26,10 @@ export class MemoryRoomManager implements IRoomManager {
     logger.info('MemoryRoomManager initialized (single-pod mode)')
   }
 
+  isReady(): boolean {
+    return true
+  }
+
   async shutdown(): Promise<void> {
     this.workflowRooms.clear()
     this.socketToWorkflow.clear()

apps/sim/socket/rooms/redis-manager.ts

@@ -96,17 +96,6 @@ export class RedisRoomManager implements IRoomManager {
     this._io = io
     this.redis = createClient({
       url: redisUrl,
-      socket: {
-        reconnectStrategy: (retries) => {
-          if (retries > 10) {
-            logger.error('Redis reconnection failed after 10 attempts')
-            return new Error('Redis reconnection failed')
-          }
-          const delay = Math.min(retries * 100, 3000)
-          logger.warn(`Redis reconnecting in ${delay}ms (attempt ${retries})`)
-          return delay
-        },
-      },
     })
 
     this.redis.on('error', (err) => {
@@ -122,12 +111,21 @@ export class RedisRoomManager implements IRoomManager {
       logger.info('Redis client ready')
       this.isConnected = true
     })
+
+    this.redis.on('end', () => {
+      logger.warn('Redis client connection closed')
+      this.isConnected = false
+    })
   }
 
   get io(): Server {
     return this._io
   }
 
+  isReady(): boolean {
+    return this.isConnected
+  }
+
   async initialize(): Promise<void> {
     if (this.isConnected) return
 

apps/sim/socket/rooms/types.ts

@@ -48,6 +48,11 @@ export interface IRoomManager {
    */
   initialize(): Promise<void>
 
+  /**
+   * Whether the room manager is ready to serve requests
+   */
+  isReady(): boolean
+
   /**
    * Clean shutdown
    */

apps/sim/socket/routes/http.ts

@@ -85,6 +85,11 @@ export function createHttpHandler(roomManager: IRoomManager, logger: Logger) {
         res.end(JSON.stringify({ error: authResult.error }))
         return
       }
+
+      if (!roomManager.isReady()) {
+        sendError(res, 'Room manager unavailable', 503)
+        return
+      }
     }
 
     // Handle workflow deletion notifications from the main API