Monitoring Services (CT 202)

LXC: monitoring

Service	Article	Port
Prometheus	prometheus	9090
Grafana	grafana	3000
Uptime Kuma	uptime kuma	3001
pve-exporter	pve exporter	9221 (internal)

node_exporter Deployment

node_exporter runs on every LXC. Deployment method depends on whether Docker is available.

LXC	Method	Komodo stack
CT 100 management	Docker	node-exporter-local
CT 200 ultrafeeder	Docker	node-exporter-ultrafeeder
CT 201 iot	Docker	node-exporter-iot
CT 202 monitoring	Docker	node-exporter-monitoring
CT 300 media-arr	Docker	node-exporter-media-arr
CT 301 media-server	Docker	node-exporter-media-server
CT 302 media-dl	Docker	node-exporter-media-dl
CT 101 pihole	Binary systemd service	-
CT 400 game-panel	Binary systemd service	-
CT 401 minecraft-wings	Binary systemd service	-

Docker stacks use network_mode: host so node_exporter binds to :9100 on each LXC's own IP.

Binary installs: /usr/local/bin/node_exporter, user node_exporter, service at /etc/systemd/system/node_exporter.service.

Dashboards

Dashboard	Grafana ID	What it shows
Node Exporter Full	1860	Per-container CPU, RAM, disk, network
Proxmox via Prometheus	10347	Proxmox host node, VMs, storage

Adding a New LXC

1. If Docker LXC: create a node-exporter-<name> Komodo stack using stacks/node-exporter/compose.yaml
2. If native LXC: install the node_exporter binary and systemd service (see CT 101 pattern)
3. Add a scrape target to /data/prometheus.yml on CT 202 and reload: curl -X POST http://localhost:9090/-/reload
4. Update the scrape target table in monitoring

Key Takeaways

• 12 active scrape targets: 10 LXCs (CT 600 not monitored) + pve-exporter + Prometheus self
• pve-exporter uses a dedicated read-only PVEAuditor token
• prometheus.yml changes take effect via HTTP reload - no container restart needed