Cilium metrics

Use Hubble metrics to monitor network traffic in your cluster.

This guide assumes that your cluster uses the Cilium CNI or that you followed the Gloo Network get started guide to install Cilium, set up Gloo Network, and deploy sample apps.

Before you begin

  1. Set up the Gloo telemetry pipeline. When you enable the pipeline, you get a few Cilium metrics by default. For more information, see View default metrics.
  2. Optional: You can enable additional Cilium metrics to gain insight into the health of the Cilium agent, Cilium operator, and your network traffic. These additional metrics are also used to feed the Cilium dashboard in Grafana.

View Cilium metrics

  1. Open the Prometheus UI

  2. Send multiple requests to the productpage page.

    for ((i=1;i<=10;i++)); do curl -vik --resolve${INGRESS_GW_IP} ; done
    for ((i=1;i<=10;i++)); do curl -vik --resolve${INGRESS_GW_IP} ; done

  3. From the Prometheus expression browser, enter the following PromQL query to see the requests the product page sent to other Bookinfo microservices.

    sum(hubble_flows_processed_total{source=~"productpage.+"}) by (destination_workload_id,verdict)
  4. Explore other queries that you run in the Prometheus expression browser to gain insight into your API Gateway.

Metric PromQL query
Successful network packet rate for a given service rate(hubble_flows_processed_total{destination_workload_id="<workload_ID>"}[5m])
Successful network packet rate between source and destination workload rate(hubble_flows_processed_total{source="<source_workload_ID>", destination_workload_id="<destination_workload_id>"}[5m])
Rate of dropped network packets to a destination workload rate(hubble_drop_total{destination_workload_id="<workload_ID>"}[5m])
Number of new processed network packets within a certain timeframe sum(increase(hubble_flows_processed_total{}[5m])) by (workload_id, destination_workload_id)