This is a LoomAI weave — a reusable slice template designed for LoomAI, the AI-assisted browser-based sandbox for designing, deploying, and managing experiments on the FABRIC testbed.

What This Weave Does

Deploys a 3-node monitoring stack. Each node is placed on a different FABRIC site, chosen at random from the set of sites that currently have enough cores, RAM, and disk to satisfy that node's requirements:

• monitor — Runs Prometheus, Grafana, and node_exporter (4 cores, 16 GB RAM, 50 GB disk)
• worker1 — Runs node_exporter (2 cores, 8 GB RAM)
• worker2 — Runs node_exporter (2 cores, 8 GB RAM)

All three nodes are connected via FABNetv4 so Prometheus can scrape metrics across the overlay network. Grafana comes pre-provisioned with a resource utilization dashboard showing CPU, memory, disk, and network metrics for all nodes. A web tunnel is automatically created so you can access the Grafana UI directly from the LoomAI Apps tab.

How to Use

1. Click Run in the LoomAI Artifacts panel and enter a slice name
2. Wait ~10–15 minutes for provisioning and software installation
3. Open the Grafana dashboard from the Apps tab (anonymous access, no login required)
4. Use the included Jupyter notebook to query Prometheus metrics and visualize CPU and network data with matplotlib
5. Click Stop to tear down the slice and free resources

Included Components

• Prometheus 2.53 — Time-series database scraping node_exporter on all nodes every 15s
• Grafana 11 — Dashboard UI with anonymous viewer access and iframe embedding enabled
• node_exporter 1.8 — Exposes hardware and OS metrics (CPU, memory, disk, network)
• Jupyter notebook — Template for querying Prometheus and graphing results with matplotlib

Installing LoomAI

LoomAI requires Docker with Compose v2. Install and run with a single command:

curl -fsSL https://raw.githubusercontent.com/fabric-testbed/loomai/main/install.sh | bash

Then open http://localhost:3000 in your browser. For more details, visit the LoomAI GitHub repository: https://github.com/fabric-testbed/loomai