Multi-Server & Scaling on Coolify — Expanding Your System

When 1 server is not enough, you need to add servers for performance, redundancy, or to separate workloads. Coolify supports managing multiple servers from a single dashboard, deploying apps to any server.

📖 Previous article: Part 10, One-Click Services

I. When Do You Need Multi-Server?

• Workload separation: Server A runs production apps, Server B runs databases, Server C runs monitoring
• Scaling: 1 server runs out of RAM/CPU → add more servers
• Geographic distribution: Servers in multiple regions → reduced latency
• Build separation: Dedicated build server (powerful CPU) to avoid impacting production
• High availability: Main server down → failover to backup server

II. Adding Remote Server

Requirements

• New server must run Linux (Ubuntu, Debian, CentOS, Alpine…)
• SSH access (port 22 or custom port)
• Root access or user with sudo
• Open ports: 80, 443 (for web traffic)
• No need to install Docker beforehand: Coolify installs it automatically

Steps to add server

1. Go to Servers → click “+ Add”
2. Fill in information:
   • Name: Server name (e.g., “production-02”)
   • IP Address: IP or hostname
   • Port: SSH port (default 22)
   • User: root (or user with sudo)
3. Choose SSH Key (use existing Coolify key or add new key)
4. Click “Validate Server”
5. Coolify SSH connects → checks → installs Docker if not present

ℹ️ Before validation, copy Coolify public key (displayed on add server page) to ~/.ssh/authorized_keys on the new server.

III. Deploy App on Remote Server

After adding servers, when creating new resources, Coolify asks which server to deploy to:

1. New Resource → choose type (App/Database/Service)
2. “Select Server” step → choose target server
3. Choose Destination (Docker network on that server)
4. Configure and Deploy as usual

IV. Dedicated Build Server

Building Docker images consumes CPU and RAM. On production servers, building can reduce performance of running apps.

Solution: separate build to dedicated server:

1. Add new server (powerful CPU, lots of RAM) → mark as “Build Server”
2. In resource config → choose build server
3. Coolify builds image on build server → pushes to production server
4. Production server only runs containers: no build resource consumption

V. Docker Swarm Mode

Coolify supports Docker Swarm for clustering, running containers across multiple nodes with automatic load balancing.

Setup Swarm

1. Go to server → “Swarm” tab
2. Click “Init Swarm” on main server (manager node)
3. Add other servers → click “Join Swarm” (worker nodes)
4. When deploying, Swarm automatically distributes containers across nodes

⚠️ Docker Swarm mode is an experimental feature on Coolify. For most use cases, multi-server without Swarm, deploying separately on each server is sufficient.

VI. Resource Limits

Limit CPU and RAM for each container:

• Go to resource → “Advanced” tab
• Set CPU limit (e.g., 1.5 cores)
• Set Memory limit (e.g., 512MB)
• Set Memory reservation (guaranteed minimum)

💡 Always set memory limits for production containers. If not, one app with memory leak can consume all server RAM, affecting all other apps.

VII. Rolling Updates

Coolify supports zero-downtime deployment:

1. When deploying new version → Coolify starts new container
2. New container must pass health check first
3. When healthy → Traefik switches traffic to new container
4. Old container is stopped and removed

Result: users see no downtime, even when deploying new versions.

VIII. Summary

Multi-server on Coolify allows scaling from a small VPS to multi-server infrastructure, all managed from one dashboard. Dedicated build servers, resource limits, rolling updates – sufficient for most production workloads. Final article: Best practices and troubleshooting.

👉 Next article: Part 12, Production Best Practices