Deployment

Mimir can be deployed in several ways depending on your needs. Docker is the recommended approach for production deployments.

Building the Docker Image

First, build the Docker image that you'll deploy to your platform:

Build the Image

docker build -t mimir-rag:latest .

Or use the Makefile:

make docker-build

Test Locally (Optional)

Before deploying, you can test the image locally:

docker run --rm \
  -p 3000:3000 \
  --env-file .env \
  mimir-rag:latest

Platform Deployment

Once you have built the Docker image, deploy it to your chosen platform.

Hetzner Deployment

Hetzner is a popular choice for deploying Mimir due to its simplicity and cost-effectiveness.

Quick Setup

1. Create a Hetzner Cloud Server:

   • Choose Ubuntu 22.04 or later
   • Recommended: 2 vCPU, 4GB RAM (or higher for production)
   • Select your preferred datacenter location

2. SSH into your server:

   ssh root@your-server-ip

3. Install Docker:

   curl -fsSL https://get.docker.com -o get-docker.sh
   sh get-docker.sh

4. Clone, build, and configure:

   git clone <your-repo-url>
   cd mimir/mimir-rag
   cp .env.example .env
   # Edit .env with your configuration
   
   # Build the Docker image
   docker build -t mimir-rag:latest .

5. Deploy the image:

   docker run -d \
     --name mimir-rag \
     -p 3000:3000 \
     --env-file .env \
     --restart unless-stopped \
     mimir-rag:latest

That's it! Your Mimir instance is now running on Hetzner.

Using Docker Compose (Optional)

For easier management, you can use Docker Compose:

# docker-compose.yml
version: '3.8'
services:
  mimir-rag:
    build: .
    ports:
      - "3000:3000"
    env_file:
      - .env
    restart: unless-stopped

Then run:

docker-compose up -d

Firewall Configuration

Make sure to open port 3000 (or your chosen port) in Hetzner's firewall:

• Go to your server's firewall settings in Hetzner Cloud Console
• Add a rule for TCP port 3000 (or your port)
• Allow traffic from your IP or 0.0.0.0/0 for public access

Reverse Proxy (Recommended)

For production, use Nginx or Caddy as a reverse proxy:

Nginx example:

server {
    listen 80;
    server_name your-domain.com;

    location / {
        proxy_pass http://localhost:3000;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
    }
}

Self-Hosted Deployment

Deploy Mimir on your own infrastructure for full control.

Requirements

• Node.js 20 or later
• Access to Supabase database
• LLM provider API keys
• Process manager (PM2 recommended)

Setup Steps

1. Clone and Install:

git clone <your-repo-url>
cd mimir-rag
npm install

2. Configure:

cp .env.example .env
# Edit .env with your configuration

3. Set Up Database:

make setup-db DB_URL=postgresql://user@host:5432/db DB_PASSWORD=secret

4. Start with PM2:

npm install -g pm2
pm2 start npm --name "mimir-rag" -- start
pm2 save
pm2 startup

Process Management

Start: pm2 start mimir-rag
Stop: pm2 stop mimir-rag
Restart: pm2 restart mimir-rag
View Logs: pm2 logs mimir-rag

Production Checklist

Before deploying to production:

• Generate a secure API key
• Set up Supabase with proper security settings
• Configure GitHub webhook secret (if using webhooks)
• Set appropriate similarity threshold and match count
• Configure exclude patterns to avoid indexing test files
• Set up monitoring and logging
• Configure backup strategy for your database

Monitoring

Health Check

Monitor your deployment with the health endpoint:

curl https://your-deployment.com/health

Expected response:

{
  "status": "ok",
  "ingestionBusy": false
}

Scaling

Mimir is stateless and can be scaled horizontally:

• Run multiple instances behind a load balancer
• All instances share the same Supabase database
• Use sticky sessions if needed (not required for most use cases)