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

  1. pinger_ovs_up
  2. pinger_ovs_down
  3. pinger_ovn_controller_up
  4. pinger_ovn_controller_down
  5. pinger_dns_healthy
  6. pinger_dns_unhealthy
  7. pinger_dns_latency_ms
  8. pinger_pod_ping_latency_ms
  9. pinger_pod_ping_lost_total
  10. pinger_node_ping_latency_ms
  11. 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.

alt text

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

alt text