Interactive architecture

Grafana Alloy Fleet and Alerting Flow

Click each node to see its role, then switch scenarios to watch metrics, logs, and alerting move through the stack.

Read blog
remote configscrape / tailPromQLlocal PromQLLogQLalertsalertsalertsalert stateFleet ManagementAlloy collectorsApps / nodesMimir / PrometheusLokiPrometheus rulesMimir RulerLoki RulerAlertmanagerGrafana
Alerting Pipeline

Prometheus rules, Mimir Ruler, and Loki Ruler evaluate conditions and send firing alerts to Alertmanager.

1Prometheus catches local failures
2Mimir Ruler evaluates global PromQL
3Loki Ruler evaluates LogQL events
4Alertmanager groups and routes alerts
5Grafana shows alert state
Rule placement

Prometheus rules vs Mimir Ruler

Prometheus rules excel

Local cluster health, bootstrap alerts, scrape health, and alerts that must work even if central remote-write is unhealthy.

Mimir Ruler excels

Global SLOs, tenant-wide recording rules, multi-cluster service alerts, and centrally governed metric alerting.

Run both local and central alerting: Prometheus for local/bootstrap alerts, Mimir Ruler for global metric alerts, Loki Ruler for log-native alerts.