Monitoring

What is collected, where it is displayed, and what alerts on it.

• :material-server-network: Collectors

  Installing Alloy on a host, by Ansible or by hand.

• :material-docker: Container metrics

  Why cAdvisor needs root, and how it fails silently without it.

• :material-view-dashboard: Dashboards

  The three provisioned dashboards and what each panel is for.

• :material-bell-alert: Alerting

  26 rules, the Matrix relay, and how to test the chain.

Coverage at a glance

Signal	Source	Where it lands
CPU, memory, disk, network, load	prometheus.exporter.unix	System Overview, VM Fleet Overview
systemd unit state	unix exporter, systemd collector	Services and Logs
Pending packages, reboot required	textfile collector	VM Fleet Overview
Container CPU, memory, restarts, health	prometheus.exporter.cadvisor	Services and Logs
Container update results	textfile collector	alerts only
systemd journal	loki.source.journal	Services and Logs
Container stdout/stderr	loki.source.docker	Services and Logs

What is deliberately not collected

• The Proxmox hypervisor itself. Its guests are monitored; the host is not. Host-level CPU, RAM, disk and ZFS pressure on the machine everything runs on is invisible here.
• OPNsense. Would need the os-node_exporter plugin, scraped rather than pushed.
• The central LXC being down. Grafana dies with it. Needs an external dead-man's-switch.

Label conventions

Every series carries these, applied by Alloy as external labels:

Label	Source	Notes
host	MONITOR_HOSTNAME	never rename, it forks history
role	MONITOR_ROLE	server, workstation, pi, router, service, central
job	set by the exporter	integrations/unix, integrations/cadvisor, alloy

role is genuinely useful in rules. fleet-container-disappeared excludes role="workstation" because a dev box starts and destroys throwaway containers constantly, and every one of them would otherwise fire.