Redis Pub/Sub for Real-Time Notifications Setup

With 10,000 WebSocket connections, each server instance only knows about its own clients. Broadcasting a real-time notification (order status update) to everyone requires a unified broker for horizontal scaling. We solve this with Redis Pub/Sub: a message is published to a channel, all active subscribers receive it. A fire-and-forget mechanism with <1 ms latency. The broker does not store history – suitable for short notifications, cache invalidation, state synchronization. Our setup for real-time notifications leverages Redis Pub/Sub for horizontal scaling. Below: configuration, integration with Node.js and Laravel, pitfalls.

How to Choose Between Pub/Sub and Redis Streams?

Pub/Sub solves the broadcast problem: one publisher, multiple subscribers, all receive identical messages simultaneously. If you need guaranteed delivery, history, or consumer groups – use Redis Streams. Pub/Sub is 2–3 times faster than Redis Streams for broadcast distribution, but does not guarantee delivery. Summary table:

How to Avoid Message Loss During Subscriber Disconnection?

A straightforward way to protect against message loss is to duplicate recent events into a separate Redis key with TTL (e.g., last_event:user:42). When a client connects, it first retrieves missed events via REST API, then switches to Pub/Sub. For critical data (financial transactions), use Redis Streams with consumer groups. This approach reduces losses to zero and saves up to 30% resources on reprocessing.

Basic Redis Configuration

Redis supports Pub/Sub out of the box. Recommended redis.conf parameters:

# Memory limit maxmemory 512mb maxmemory-policy allkeys-lru # Number of databases databases 16 # Disable persistence for pure Pub/Sub save "" appendonly no 

For production, use Redis Sentinel or Cluster. Regular Pub/Sub in Cluster is limited to a shard – with Redis 7+, use sharded Pub/Sub (SPUBLISH/SSUBSCRIBE).

Integration with Node.js and Socket.io

Use the ioredis library for two separate connections (publisher and subscriber). Example in TypeScript:

import Redis from 'ioredis'; const publisher = new Redis({ host: 'redis', port: 6379 }); const subscriber = new Redis({ host: 'redis', port: 6379 }); subscriber.subscribe('notifications:user:*', (err, count) => { if (err) throw err; console.log(`Subscribed to ${count} channels`); }); subscriber.on('pmessage', (pattern, channel, message) => { const userId = channel.split(':')[2]; const payload = JSON.parse(message); broadcastToUser(userId, payload); }); async function notifyUser(userId: string, event: object) { const channel = `notifications:user:${userId}`; const count = await publisher.publish(channel, JSON.stringify(event)); return count; } 

For WebSocket, use Socket.io with Redis Adapter:

import { createServer } from 'http'; import { Server } from 'socket.io'; import { createAdapter } from '@socket.io/redis-adapter'; import { createClient } from 'redis'; const httpServer = createServer(); const io = new Server(httpServer); const pubClient = createClient({ url: 'redis://redis:6379' }); const subClient = pubClient.duplicate(); await Promise.all([pubClient.connect(), subClient.connect()]); io.adapter(createAdapter(pubClient, subClient)); httpServer.listen(3000); 

Redis Adapter uses Pub/Sub: when io.to('room').emit() is called, the command is published to a Redis channel, and all server instances broadcast it to the room's clients.

Integration with Laravel

In Laravel, sending a Pub/Sub message looks like this:

use Illuminate\Support\Facades\Redis; Redis::publish('notifications:user:'.$userId, json_encode([ 'type' => 'order.status_changed', 'orderId' => $order->id, 'status' => $order->status, 'timestamp' => now()->toISOString(), ])); 

For receiving messages on the client, use Laravel Echo Server or Soketi.

Process of Setup (Step by Step)

1. Design channel scheme: commonly use patterns entity:action:user_id.
2. Configure Redis: set maxmemory, disable AOF for pure Pub/Sub.
3. Implement event publishing in the backend (Node.js, Laravel, Django).
4. Connect WebSocket server (Socket.io, Soketi) with Redis Adapter.
5. Develop fallback request mechanism for missed events (via REST).
6. Load test: ensure instantaneous_ops_per_sec does not exceed 5,000.

Monitoring

Use Redis CLI commands:

redis-cli PUBSUB CHANNELS "*" redis-cli PUBSUB NUMSUB notifications:user:42 redis-cli PUBSUB NUMPAT 

In Prometheus with Redis Exporter, monitor instantaneous_ops_per_sec. Growth above 5,000 ops/s under Pub/Sub load signals the need for optimization (reduce message size, increase number of channels).

Limitations and Alternatives

Message Loss on Reconnect

A subscriber that disconnects temporarily will miss messages sent during that time. For critical notifications – Redis Streams or store recent events.

No Delivery Confirmation

PUBLISH returns the number of receivers but does not guarantee processing. For at-least-once – use a queue (RabbitMQ, Redis Streams).

CPU Load with Many Patterns

PSUBSCRIBE matches every message against all patterns. With 10,000+ patterns, latency increases by 30–50%.

What's Included in the Work

Implementing Pub/Sub can reduce WebSocket infrastructure costs by up to 40%. For a typical application with 10,000 concurrent users, this setup reduces monthly WebSocket costs by approximately $1,500. A typical Redis instance for Pub/Sub costs around $30-50 per month on cloud providers like AWS or DigitalOcean. Our setup for real-time notifications leverages Redis Pub/Sub, Node.js ioredis, and Socket.io Redis adapter for seamless horizontal scaling. With over 6 years of experience and 15+ successful projects, we provide expert Redis consulting.

• Designing channel and pattern schema for business logic.
• Configuring Redis (configuration, clustering, monitoring).
• Integrating Pub/Sub with the backend (Node.js, Laravel, Django – according to your stack).
• Connecting WebSocket server (Socket.io, Soketi, Laravel Echo).
• Developing message loss prevention (fallback requests).
• Deployment and operations documentation.
• Repository and infrastructure access transfer.
• Support for 2 weeks after delivery.

We have been working with Redis for over 6 years and have implemented 15+ Pub/Sub projects for e-commerce and fintech clients. We use production configurations with Sentinel and Cluster, ensuring fault tolerance. Contact us to set up Redis Pub/Sub for your project – get a consultation within 1 day. Reach out to us for implementation and scaling.