Most teams evaluating our systems already run Prometheus and Grafana for infrastructure monitoring. Rather than ask you to adopt yet another dashboard, our edge and on-site software exposes a Prometheus-compatible metrics endpoint so the data lands where your operators already look.
What gets exported
Each edge node and on-site aggregation server publishes a /metrics endpoint that Prometheus can
scrape on your own interval. Typical series include:
- Node health: CPU, memory, thermal, and storage headroom per ruggedised camera node
- Inference throughput: frames processed, events raised, and end-to-end detection latency
- Fleet state: nodes online, last-seen timestamps, and model version per device
- Link quality: uplink availability for bandwidth-constrained and unattended sites
Wiring it up
Point a Prometheus scrape job at the on-site server, which federates metrics from every node behind the site perimeter, so raw video never leaves, but operational telemetry does. Import our starter Grafana dashboard JSON for fleet overview and per-node drill-down panels, then layer your own alerting rules through Alertmanager.
Because the integration is pull-based and read-only, it fits cleanly inside air-gapped and on-premises deployments common to railways and defence, and adds no new data-egress path. If you want to see it against your own monitoring stack, talk to us.