OpenWebUI + Ollama Docker Compose Setup

This Docker Compose configuration provides an AI stack with three main services:

• Ollama: AI model server with NVIDIA GPU support
• Open WebUI: Web interface for interacting with Ollama models
• Uptime Kuma: Monitoring solution for service health

Requirements

• Docker and Docker Compose
• NVIDIA GPU with compatible drivers
• NVIDIA Container Toolkit installed

Usage

1. Start the services:

   just start

2. Access the services:

   • Open WebUI: http://localhost:8080
   • Uptime Kuma: http://localhost:3001

Uptime Kuma Configuration

After starting the services, configure Uptime Kuma to monitor both Ollama and Open WebUI:

1. Access Uptime Kuma at http://localhost:3001
2. Create a new account during first login (not really yet)
3. Add monitors for both services:

Basic Health Check for Open WebUI

1. Click "Add New Monitor"
2. Configure:
   • Monitor Type: HTTP(s)
   • Name: Open WebUI Health
   • URL: http://open-webui:8080/health
   • Monitoring Interval: 60 seconds
   • Retry Count: 3

Basic Health Check for Ollama

1. Click "Add New Monitor"
2. Configure:
   • Monitor Type: HTTP(s)
   • Name: Ollama Health
   • URL: http://ollama:11434/api/health
   • Monitoring Interval: 60 seconds
   • Retry Count: 3

Volume Data

All services use persistent volumes:

• Ollama data: /root/.ollama
• Open WebUI data: /app/backend/data
• Uptime Kuma data: /app/data

Additional Configuration

For more advanced monitoring options, including Model Connectivity and Response Testing, refer to the Open WebUI Monitoring Documentation.