Back to Projects
Observability

Observability Stack

COMPLETE SURFACE: BARE METAL

Prometheus and Grafana watching the Docker host that now runs every self-hosted service. Originally covered a k3s cluster too, before those nodes died.

Metrics
Prometheus
Dashboards
Grafana
Coverage
Docker host
Next
Loki

Setup

Prometheus scraping node_exporter. It originally covered three Lenovo k3s nodes plus the Beelink; the Lenovos are gone now (see Homelab Kubernetes Cluster), so it’s just the Docker host. Grafana for visualization. Basic alerting for uptime and disk.

What’s Missing

Log aggregation. Loki is the obvious next step, but log retention at homelab scale needs a real storage plan first. Alertmanager is also on the list. Right now alerts exist in Grafana but nothing acts on them.

Stack

Prometheus · Grafana · node_exporter