Metrics and alerting

The core provides a module named metrics that collects metrics from the nodes and sends alerts to the Nethesis portal. The module is installed by the create-cluster action.

The module includes the following services:

• Prometheus
• Alertmanager
• Grafana
• alert-proxy

Key points:

• Single instance running on the leader node
• Monitors all cluster nodes automatically
• Removed if the leader node becomes a worker
• Prometheus: port 9091
• Alertmanager: port 9093
• alert-proxy: port 9095
• Grafana: port 3000 (disabled by default, enabled with Traefik route)

Configuration:

• Prometheus and Alertmanager configurations are recreated on Prometheus restart
• Module restarts on node addition/removal
• alert-proxy restarts on subscription change to send alerts to Nethesis portals

If a subscription is enabled, alerts are sent to Nethesis portals by default.

Please refer to the module README for details on alert configuration and customization.