n8n queue mode with Redis

In production, running n8n in queue mode prevents a slow workflow (a 30s LLM call, for example) from blocking the entire webhook queue.

Architecture

Webhook ──▶ n8n main (HTTP) ──▶ Redis (queue) ──▶ n8n worker(s) ──▶ Postgres

• main: receives the webhook, enqueues it and returns 200 immediately
• worker: pulls from the queue, executes and persists to Postgres
• Redis: queue broker (BullMQ under the hood)

Prerequisites

• Postgres 14+ (recommended: container or external service)
• Redis 7+
• VPS with 4+ vCPU and 8 GB of RAM

Minimal docker-compose.yml

services:
  postgres:
    image: postgres:16
    environment:
      POSTGRES_USER: n8n
      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD}
      POSTGRES_DB: n8n
    volumes:
      - pg_data:/var/lib/postgresql/data
    restart: unless-stopped

  redis:
    image: redis:7-alpine
    restart: unless-stopped

  n8n-main:
    image: docker.n8n.io/n8nio/n8n
    environment:
      EXECUTIONS_MODE: queue
      QUEUE_BULL_REDIS_HOST: redis
      DB_TYPE: postgresdb
      DB_POSTGRESDB_HOST: postgres
      DB_POSTGRESDB_USER: n8n
      DB_POSTGRESDB_PASSWORD: ${POSTGRES_PASSWORD}
      DB_POSTGRESDB_DATABASE: n8n
      N8N_HOST: n8n.exemplo.com
      WEBHOOK_URL: https://n8n.exemplo.com/
      N8N_PROTOCOL: https
      N8N_ENCRYPTION_KEY: ${N8N_ENCRYPTION_KEY}
    ports:
      - "5678:5678"
    depends_on: [postgres, redis]
    restart: unless-stopped

  n8n-worker:
    image: docker.n8n.io/n8nio/n8n
    command: worker
    environment:
      EXECUTIONS_MODE: queue
      QUEUE_BULL_REDIS_HOST: redis
      DB_TYPE: postgresdb
      DB_POSTGRESDB_HOST: postgres
      DB_POSTGRESDB_USER: n8n
      DB_POSTGRESDB_PASSWORD: ${POSTGRES_PASSWORD}
      DB_POSTGRESDB_DATABASE: n8n
      N8N_ENCRYPTION_KEY: ${N8N_ENCRYPTION_KEY}
    depends_on: [postgres, redis]
    restart: unless-stopped

volumes:
  pg_data:

Scale workers

For more throughput, scale only the workers:

docker compose up -d --scale n8n-worker=4

Each worker takes 1 job at a time by default. For light jobs (no LLM), raise N8N_CONCURRENCY=10 per worker.

How to tell it is working

Look for this log in the workers on startup:

n8n Task Broker is now accepting connections on port 5679
n8n worker is now ready
- version: 1.x.x
- concurrency: 10

And in main:

Editor is now accessible via:
https://n8n.exemplo.com/

Monitoring

We recommend exposing BullMQ metrics via /metrics (Prometheus) and charting them in Grafana:

• Active / waiting / completed jobs
• Average execution time per workflow
• Failures per hour

Next steps

• Public webhooks
• n8n + EvolutionAPI + OpenAI