prometheus.md

Pinger makes network requests between pods/nodes/services/dns to test the connectivity in the cluster and expose metrics in Prometheus format.

Prometheus Integration

Pinger exposes metrics at :8080/metrics, it will show following metrics

pinger_ovs_up
pinger_ovs_down
pinger_ovn_controller_up
pinger_ovn_controller_down
pinger_dns_healthy
pinger_dns_unhealthy
pinger_dns_latency_ms
pinger_pod_ping_latency_ms
pinger_pod_ping_lost_total
pinger_node_ping_latency_ms
pinger_node_ping_lost_total

Kube-OVN-Controller expose metrics at 10660/metrics, it will show controller runtime metrics.

You can use kube-prometheus to scrape the metrics. The related ServiceMonitor yaml can be found here

Grafana Dashboard

Pinger grafana dashboard config can be found here.

Kube-OVN-Controller grafana dashboard config can be found here

Kube-OVN-CNI grafana dashboard config can be found here