Observability¶
Metrics, logs, dashboards, alerting, and uptime monitoring.
All services run on the k3s-cluster (monitoring namespace).
Grafana¶
monitoring
Dashboards and visualisation. Connects to Prometheus, Loki, and InfluxDB to provide unified monitoring views across the whole stack.
Documentation · values.sops.yaml
Prometheus Stack¶
monitoring
Metrics collection and alerting. Deployed via kube-prometheus-stack, includes Prometheus, Alertmanager, kube-state-metrics, and node exporters.
Documentation · values.sops.yaml
Loki¶
monitoring
Log aggregation. Collects and indexes logs from all pods and ships them to Grafana for querying with LogQL.
Alloy¶
monitoring
OpenTelemetry collector and telemetry pipeline. Scrapes metrics, tails logs, and forwards traces to the appropriate backends (Prometheus, Loki).
Uptime Kuma¶
monitoring
Service uptime monitoring. Checks HTTP endpoints, TCP ports, and DNS records at configurable intervals and sends alerts on downtime.
ntfy¶
monitoring
Self-hosted push notification server. Receives alerts from Prometheus Alertmanager, Uptime Kuma, and PagerDuty and delivers them to mobile devices.
Documentation · values.sops.yaml
Speedtest Exporter¶
monitoring
Telegraf instance with an exec input that runs periodic internet speed tests and exposes results as Prometheus metrics for Grafana dashboards.
Unpoller¶
monitoring
Exports UniFi controller metrics (clients, traffic, device health) to Prometheus for visualisation in Grafana.
Cluster Heartbeat¶
monitoring
Custom heartbeat job that sends a periodic ping to healthchecks.io. If the ping stops, healthchecks.io fires a PagerDuty alert — confirming the cluster is alive end-to-end.
OctoTrack¶
monitoring
Octopus Energy electricity usage dashboard. Pulls consumption data from the Octopus Energy API and displays it in Grafana.