AINFRA is a monitoring project that has reached its alpha version.
- Git installed on your system
- Docker and Docker Compose installed on your system
- OpenAI API key (Required for LLM service)
- For standard devices: Glances running on target systems
-
Clone the repository:
git clone https://github.com/n1kozor/AINFRA.git cd AINFRA -
Build and start the Docker containers:
docker build docker-compose up -d
-
For standard devices (Windows, Linux, Mac), install Glances ON THE TARGET DEVICE:
docker run -d --restart="always" -p 61208-61209:61208-61209 -e GLANCES_OPT="-w" \ -v /var/run/docker.sock:/var/run/docker.sock:ro \ -v /run/user/1000/podman/podman.sock:/run/user/1000/podman/podman.sock:ro \ --pid host nicolargo/glances:latest
Currently, the LLM service only works with OpenAI. During the first startup, you'll need to configure the service:
- Enter your OpenAI API key in the settings section
- The system will retrieve the available models that you can access with your API key
- Select which model you would like to use from the list of available models
- Requires Glances to be running
- All statistics are transferred to the MCP server
- LLM can be used to query any system information
- Create custom plugins to add any networked device
- Successfully tested with: ESXi, TVs, lab equipment, Synology NAS, Proxmox, Fritz!Box routers
- Supports both querying and controlling devices
- LLM can understand and execute operations defined in plugins
- Device Management: Add and monitor devices on the Devices page
- Chat Interface: Communicate with your devices through the chat interface
- Dashboard: Click on the floating "soul" icon to generate network summaries
- Sensors: Create RAM and CPU monitors (currently for standard devices only)
- New devices require 30-50 seconds before becoming available due to connectivity checks
- Auto-refresh functionality needs improvement; manual page refreshes often required
- Plugins can only be added in JSON format
- Device filtering is not yet implemented
- Ollama integration is still in development
- Additional sensor types (processes monitoring, etc.)
- Sensor support for custom devices
- Custom agent for standard devices
- Enhanced dynamic interface for plugin-based devices