# Node-Exporter

Node-Exporter exposes host metrics (CPU, RAM, disk, network) to Prometheus. One instance runs per host inside its own container.

**Hosts:** Dankserver + Dankclaw — **Docs:** https://prometheus.io/docs/guides/node-exporter

## Facts (from research)

| Item | Value |
|------|-------|
| Image | `prom/node-exporter:latest` |
| Port | 9100 (host) |
| Network | `host` mode |
| Config | `--path.procfs=/host/proc`, `--path.sysfs=/host/sys`, `--path.rootfs=/rootfs` |

## Configuration

```yaml
services:
  node-exporter:
    image: prom/node-exporter:latest
    restart: always
    network_mode: host
    pid: host
    volumes:
      - /proc:/host/proc:ro
      - /sys:/host/sys:ro
      - /:/rootfs:ro
    command:
      - --path.procfs=/host/proc
      - --path.sysfs=/host/sys
      - --path.rootfs=/rootfs
```

## Prometheus Targets

Prometheus scrapes each host's node-exporter directly:

| Host | Target |
|------|--------|
| Dankserver | `192.168.0.235:9100` |
| Dankclaw | `192.168.0.143:9100` |

These scrape jobs live in the Prometheus config; see [[Prometheus]] and [[Observability]].

## Gotchas

- Requires `network_mode: host` (or a published port) so Prometheus can reach 9100.
- The `/proc`, `/sys`, and root mounts must be `read-only` (`:ro`).
- Do not mount the same paths on both hosts' containers; each node-exporter reflects only its own host.

## See Also

- [[Prometheus]] — the scraper
- [[Observability]] — the full stack