Horizontal Scaling

Overview

Sockudo supports horizontal scaling across multiple nodes using Redis, Redis Cluster, or NATS as message brokers. Scale to handle 100K+ concurrent connections with proper load balancing.

Architecture

┌─────────────────┐
│   Load Balancer │  (Nginx/HAProxy)
│  Sticky Sessions│
└────────┬────────┘
         │
    ┌────┴─────┬──────────┬──────────┐
    │          │          │          │
┌───▼──┐   ┌──▼───┐  ┌───▼──┐   ┌──▼───┐
│Node 1│   │Node 2│  │Node 3│   │Node N│
└───┬──┘   └──┬───┘  └───┬──┘   └──┬───┘
    │         │          │         │
    └─────────┴──────────┴─────────┘
              │
         ┌────▼─────┐
         │  Redis   │  (Adapter)
         │ Cluster  │
         └──────────┘

Choosing an Adapter

Redis (Recommended)

Best for: Most production deployments, 2-10 nodes Pros:

Simple setup
Proven reliability
Built-in persistence
Low latency (~1-2ms)

Cons:

Single point of failure (use Sentinel for HA)
Limited to single master

environment:
  ADAPTER_DRIVER: "redis"
  DATABASE_REDIS_HOST: "redis"
  DATABASE_REDIS_PORT: "6379"
  DATABASE_REDIS_PASSWORD: "${REDIS_PASSWORD}"

Redis Cluster

Best for: Large deployments, 10+ nodes, 100K+ connections Pros:

Horizontal scaling
Automatic sharding
High availability
No single point of failure

Cons:

More complex setup
Higher operational overhead

environment:
  ADAPTER_DRIVER: "redis-cluster"
  REDIS_CLUSTER_NODES: "redis-1:6379,redis-2:6379,redis-3:6379,redis-4:6379,redis-5:6379,redis-6:6379"

NATS

Best for: Cloud-native deployments, Kubernetes, multi-region Pros:

Built for distributed systems
Excellent multi-region support
Low latency (<1ms)
Native clustering

Cons:

No persistence by default
Less common than Redis

environment:
  ADAPTER_DRIVER: "nats"
  NATS_SERVERS: "nats://nats-1:4222,nats://nats-2:4222,nats://nats-3:4222"

Multi-Node Setup with Redis

Docker Compose Configuration

services:
  # Shared Redis for all nodes
  redis:
    image: redis:7-alpine
    command: >
      redis-server
      --requirepass ${REDIS_PASSWORD}
      --appendonly yes
      --maxmemory 1gb
      --maxmemory-policy allkeys-lru
    volumes:
      - redis-data:/data
    networks:
      - sockudo-network
    deploy:
      resources:
        limits:
          memory: 1G
          cpus: '2.0'

  # Sockudo Node 1
  sockudo-node1:
    build: .
    environment:
      ADAPTER_DRIVER: "redis"
      CACHE_DRIVER: "redis"
      QUEUE_DRIVER: "redis"
      RATE_LIMITER_DRIVER: "redis"
      
      DATABASE_REDIS_HOST: "redis"
      DATABASE_REDIS_PASSWORD: "${REDIS_PASSWORD}"
      
      INSTANCE_PROCESS_ID: "sockudo-node1"
      HOST: "0.0.0.0"
      PORT: "6001"
      METRICS_PORT: "9601"
    ports:
      - "6001:6001"
      - "9601:9601"
    depends_on:
      redis:
        condition: service_healthy
    networks:
      - sockudo-network

  # Sockudo Node 2
  sockudo-node2:
    build: .
    environment:
      ADAPTER_DRIVER: "redis"
      CACHE_DRIVER: "redis"
      QUEUE_DRIVER: "redis"
      RATE_LIMITER_DRIVER: "redis"
      
      DATABASE_REDIS_HOST: "redis"
      DATABASE_REDIS_PASSWORD: "${REDIS_PASSWORD}"
      
      INSTANCE_PROCESS_ID: "sockudo-node2"
      HOST: "0.0.0.0"
      PORT: "6002"
      METRICS_PORT: "9602"
    ports:
      - "6002:6002"
      - "9602:9602"
    depends_on:
      redis:
        condition: service_healthy
    networks:
      - sockudo-network

  # Sockudo Node 3
  sockudo-node3:
    build: .
    environment:
      ADAPTER_DRIVER: "redis"
      CACHE_DRIVER: "redis"
      QUEUE_DRIVER: "redis"
      RATE_LIMITER_DRIVER: "redis"
      
      DATABASE_REDIS_HOST: "redis"
      DATABASE_REDIS_PASSWORD: "${REDIS_PASSWORD}"
      
      INSTANCE_PROCESS_ID: "sockudo-node3"
      HOST: "0.0.0.0"
      PORT: "6003"
      METRICS_PORT: "9603"
    ports:
      - "6003:6003"
      - "9603:9603"
    depends_on:
      redis:
        condition: service_healthy
    networks:
      - sockudo-network

  # Load Balancer
  nginx:
    image: nginx:alpine
    ports:
      - "80:80"
      - "443:443"
    volumes:
      - ./nginx/nginx-lb.conf:/etc/nginx/nginx.conf:ro
      - ./nginx/ssl:/etc/nginx/ssl:ro
    depends_on:
      - sockudo-node1
      - sockudo-node2
      - sockudo-node3
    networks:
      - sockudo-network

volumes:
  redis-data:

networks:
  sockudo-network:
    driver: bridge

Load Balancing

Nginx Configuration

Create nginx/nginx-lb.conf:

upstream sockudo_backend {
    # Sticky sessions using IP hash
    ip_hash;
    
    server sockudo-node1:6001 max_fails=3 fail_timeout=30s;
    server sockudo-node2:6002 max_fails=3 fail_timeout=30s;
    server sockudo-node3:6003 max_fails=3 fail_timeout=30s;
}

server {
    listen 80;
    server_name ws.your-domain.com;
    
    # Redirect to HTTPS
    return 301 https://$server_name$request_uri;
}

server {
    listen 443 ssl http2;
    server_name ws.your-domain.com;
    
    ssl_certificate /etc/nginx/ssl/fullchain.pem;
    ssl_certificate_key /etc/nginx/ssl/privkey.pem;
    ssl_protocols TLSv1.2 TLSv1.3;
    ssl_ciphers HIGH:!aNULL:!MD5;
    
    location / {
        proxy_pass http://sockudo_backend;
        proxy_http_version 1.1;
        
        # WebSocket support
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection "upgrade";
        
        # Headers
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
        proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
        proxy_set_header X-Forwarded-Proto $scheme;
        
        # Timeouts
        proxy_connect_timeout 7d;
        proxy_send_timeout 7d;
        proxy_read_timeout 7d;
    }
    
    # Health check endpoint
    location /health {
        access_log off;
        return 200 "OK";
    }
}

HAProxy Configuration

Alternatively, use HAProxy:

global
    log /dev/log local0
    maxconn 50000

defaults
    log global
    mode http
    option httplog
    timeout connect 5s
    timeout client 7d
    timeout server 7d

frontend websocket
    bind *:80
    bind *:443 ssl crt /etc/haproxy/certs/sockudo.pem
    default_backend sockudo_nodes

backend sockudo_nodes
    balance source  # Sticky sessions using source IP
    option httpchk GET /up
    
    server node1 sockudo-node1:6001 check
    server node2 sockudo-node2:6002 check
    server node3 sockudo-node3:6003 check

Sticky Sessions

Why Sticky Sessions?

WebSocket connections are stateful and must remain on the same node. Use one of these methods:

IP Hash (Nginx)

upstream sockudo_backend {
    ip_hash;  # Routes same IP to same server
    server sockudo-node1:6001;
    server sockudo-node2:6002;
    server sockudo-node3:6003;
}

Pros: Simple, no cookies Cons: Not ideal for users behind NAT/proxies

backend sockudo_nodes
    balance roundrobin
    cookie SERVERID insert indirect nocache
    
    server node1 sockudo-node1:6001 check cookie node1
    server node2 sockudo-node2:6002 check cookie node2
    server node3 sockudo-node3:6003 check cookie node3

Pros: Works with NAT, better distribution Cons: Requires cookie support

Source IP (HAProxy)

backend sockudo_nodes
    balance source
    server node1 sockudo-node1:6001 check
    server node2 sockudo-node2:6002 check
    server node3 sockudo-node3:6003 check

Redis Cluster Setup

Creating Redis Cluster

services:
  redis-1:
    image: redis:7-alpine
    command: redis-server --port 6379 --cluster-enabled yes --cluster-config-file nodes.conf --cluster-node-timeout 5000 --appendonly yes
    ports:
      - "6379:6379"
    volumes:
      - redis-1-data:/data
    networks:
      - sockudo-network

  redis-2:
    image: redis:7-alpine
    command: redis-server --port 6380 --cluster-enabled yes --cluster-config-file nodes.conf --cluster-node-timeout 5000 --appendonly yes
    ports:
      - "6380:6380"
    volumes:
      - redis-2-data:/data
    networks:
      - sockudo-network

  redis-3:
    image: redis:7-alpine
    command: redis-server --port 6381 --cluster-enabled yes --cluster-config-file nodes.conf --cluster-node-timeout 5000 --appendonly yes
    ports:
      - "6381:6381"
    volumes:
      - redis-3-data:/data
    networks:
      - sockudo-network

  # Additional nodes for replicas
  redis-4:
    image: redis:7-alpine
    command: redis-server --port 6382 --cluster-enabled yes --cluster-config-file nodes.conf --cluster-node-timeout 5000 --appendonly yes
    ports:
      - "6382:6382"
    volumes:
      - redis-4-data:/data
    networks:
      - sockudo-network

  redis-5:
    image: redis:7-alpine
    command: redis-server --port 6383 --cluster-enabled yes --cluster-config-file nodes.conf --cluster-node-timeout 5000 --appendonly yes
    ports:
      - "6383:6383"
    volumes:
      - redis-5-data:/data
    networks:
      - sockudo-network

  redis-6:
    image: redis:7-alpine
    command: redis-server --port 6384 --cluster-enabled yes --cluster-config-file nodes.conf --cluster-node-timeout 5000 --appendonly yes
    ports:
      - "6384:6384"
    volumes:
      - redis-6-data:/data
    networks:
      - sockudo-network

volumes:
  redis-1-data:
  redis-2-data:
  redis-3-data:
  redis-4-data:
  redis-5-data:
  redis-6-data:

Initialize Cluster

# Create cluster with 3 masters and 3 replicas
docker exec -it redis-1 redis-cli --cluster create \
  redis-1:6379 \
  redis-2:6380 \
  redis-3:6381 \
  redis-4:6382 \
  redis-5:6383 \
  redis-6:6384 \
  --cluster-replicas 1

Configure Sockudo

environment:
  ADAPTER_DRIVER: "redis-cluster"
  CACHE_DRIVER: "redis-cluster"
  QUEUE_DRIVER: "redis-cluster"
  RATE_LIMITER_DRIVER: "redis-cluster"
  
  REDIS_CLUSTER_NODES: "redis-1:6379,redis-2:6380,redis-3:6381,redis-4:6382,redis-5:6383,redis-6:6384"

NATS Setup

NATS Cluster

services:
  nats-1:
    image: nats:latest
    command: 
      - "-js"
      - "-m"
      - "8222"
      - "--cluster"
      - "nats://0.0.0.0:6222"
      - "--routes"
      - "nats://nats-2:6222,nats://nats-3:6222"
    ports:
      - "4222:4222"
      - "8222:8222"
    networks:
      - sockudo-network

  nats-2:
    image: nats:latest
    command:
      - "-js"
      - "-m"
      - "8222"
      - "--cluster"
      - "nats://0.0.0.0:6222"
      - "--routes"
      - "nats://nats-1:6222,nats://nats-3:6222"
    ports:
      - "4223:4222"
      - "8223:8222"
    networks:
      - sockudo-network

  nats-3:
    image: nats:latest
    command:
      - "-js"
      - "-m"
      - "8222"
      - "--cluster"
      - "nats://0.0.0.0:6222"
      - "--routes"
      - "nats://nats-1:6222,nats://nats-2:6222"
    ports:
      - "4224:4222"
      - "8224:8222"
    networks:
      - sockudo-network

Configure Sockudo

environment:
  ADAPTER_DRIVER: "nats"
  NATS_SERVERS: "nats://nats-1:4222,nats://nats-2:4222,nats://nats-3:4222"

Cluster Health Monitoring

Sockudo includes built-in cluster health tracking:

{
  "adapter": {
    "cluster_health": {
      "enabled": true,
      "heartbeat_interval_ms": 10000,
      "node_timeout_ms": 30000,
      "cleanup_interval_ms": 10000
    }
  }
}

Or via environment:

environment:
  ADAPTER_CLUSTER_HEALTH_ENABLED: "true"
  ADAPTER_CLUSTER_HEALTH_HEARTBEAT_INTERVAL_MS: "10000"
  ADAPTER_CLUSTER_HEALTH_NODE_TIMEOUT_MS: "30000"
  ADAPTER_CLUSTER_HEALTH_CLEANUP_INTERVAL_MS: "10000"

Features:

Automatic node discovery
Heartbeat monitoring
Failed node detection
Automatic cleanup of stale connections

Performance Tuning

Buffer Sizing

environment:
  # Scale adapter buffer with CPU cores
  ADAPTER_BUFFER_MULTIPLIER_PER_CPU: "128"
  
  # Cleanup queue for mass disconnects
  CLEANUP_QUEUE_BUFFER_SIZE: "50000"
  CLEANUP_BATCH_SIZE: "25"

Redis Tuning

redis:
  command: >
    redis-server
    --maxmemory 2gb
    --maxmemory-policy allkeys-lru
    --tcp-backlog 511
    --timeout 0
    --tcp-keepalive 300
    --save 900 1
    --save 300 10

Database Pooling

environment:
  DATABASE_POOLING_ENABLED: "true"
  DATABASE_POOL_MIN: "4"
  DATABASE_POOL_MAX: "32"
  
  # Per-database overrides
  DATABASE_MYSQL_POOL_MIN: "4"
  DATABASE_MYSQL_POOL_MAX: "32"

Testing Multi-Node Setup

Use the provided test configuration:

# Start multi-node test environment
docker-compose -f docker-compose.multinode.yml up

# Open test client
open http://localhost:82/client/

# Connect to different nodes
# Node 1: http://localhost:6002
# Node 2: http://localhost:6003
# Load Balancer: http://localhost:82

Scaling Commands

# Scale to 5 nodes
docker-compose up -d --scale sockudo=5

# View all nodes
docker-compose ps

# View logs from all nodes
docker-compose logs -f sockudo

# Check specific node
docker-compose logs -f sockudo-node1

# Restart single node (zero downtime)
docker-compose restart sockudo-node1

Next Steps

Set up monitoring for multi-node clusters
Review production requirements
Configure Docker deployment options

Getting Started

Core Concepts

Configuration

Deployment

Advanced Features

Horizontal Scaling

Overview

Architecture

Choosing an Adapter

Redis (Recommended)

Redis Cluster

NATS

Multi-Node Setup with Redis

Docker Compose Configuration

Load Balancing

Nginx Configuration

HAProxy Configuration

Sticky Sessions

Why Sticky Sessions?

IP Hash (Nginx)

Source IP (HAProxy)

Redis Cluster Setup

Creating Redis Cluster

Initialize Cluster

Configure Sockudo

NATS Setup

NATS Cluster

Configure Sockudo

Cluster Health Monitoring

Performance Tuning

Buffer Sizing

Redis Tuning

Database Pooling

Testing Multi-Node Setup

Scaling Commands

Next Steps

Build docs developers (and LLMs) love

Getting Started

Core Concepts

Configuration

Deployment

Advanced Features

​Overview

​Architecture

​Choosing an Adapter

​Redis (Recommended)

​Redis Cluster

​NATS

​Multi-Node Setup with Redis

​Docker Compose Configuration

​Load Balancing

​Nginx Configuration

​HAProxy Configuration

​Sticky Sessions

​Why Sticky Sessions?

​IP Hash (Nginx)

​Cookie-Based (HAProxy)

​Source IP (HAProxy)

​Redis Cluster Setup

​Creating Redis Cluster

​Initialize Cluster

​Configure Sockudo

​NATS Setup

​NATS Cluster

​Configure Sockudo

​Cluster Health Monitoring

​Performance Tuning

​Buffer Sizing

​Redis Tuning

​Database Pooling

​Testing Multi-Node Setup

​Scaling Commands

​Next Steps

Build docs developers (and LLMs) love

Overview

Architecture

Choosing an Adapter

Redis (Recommended)

Redis Cluster

NATS

Multi-Node Setup with Redis

Docker Compose Configuration

Load Balancing

Nginx Configuration

HAProxy Configuration

Sticky Sessions

Why Sticky Sessions?

IP Hash (Nginx)

Cookie-Based (HAProxy)

Source IP (HAProxy)

Redis Cluster Setup

Creating Redis Cluster

Initialize Cluster

Configure Sockudo

NATS Setup

NATS Cluster

Configure Sockudo

Cluster Health Monitoring

Performance Tuning

Buffer Sizing

Redis Tuning

Database Pooling

Testing Multi-Node Setup

Scaling Commands

Next Steps