Real production N8N

Architecture

Design decisions

High availability. Two N8N instances behind Traefik, sharing Redis for queue distribution and PostgreSQL for persistent state. If the primary fails, the replica seamlessly takes over.

Security-first. All traffic enters through WireGuard VPN. Traefik terminates TLS and handles rate limiting. The entire deployment lives in a private network — no exposed ports beyond the VPN endpoint.

Observability built in. Prometheus scrapes metrics from both N8N instances. Grafana provides dashboards for workflow health, execution latency, queue depth, and resource utilization.

Automated resilience. Redis handles workflow queuing and inter-instance broadcast. PostgreSQL is clustered for data durability. Daily automated backups ensure recovery capability.