Accessing metrics

You access metrics through the Grafana UI. Grafana is the visualization tool for Prometheus.

Grafana

  1. To open Grafana, enter the following command:
  1. kubectl port-forward --namespace knative-monitoring \
  2. $(kubectl get pods --namespace knative-monitoring \
  3. --selector=app=grafana --output=jsonpath="{.items..metadata.name}") \
  4. 3000
  • This starts a local proxy of Grafana on port 3000. For security reasons, the Grafana UI is exposed only within the cluster.

  1. Navigate to the Grafana UI at http://localhost:3000.

  2. Select the Home button on the top of the page to see the list of pre-installed dashboards (screenshot below): Knative Dashboards

The following dashboards are pre-installed with Knative Serving:

  • Revision HTTP Requests: HTTP request count, latency, and size metrics per revision and per configuration
  • Nodes: CPU, memory, network, and disk metrics at node level
  • Pods: CPU, memory, and network metrics at pod level
  • Deployment: CPU, memory, and network metrics aggregated at deployment level
  • Istio, Mixer and Pilot: Detailed Istio mesh, Mixer, and Pilot metrics
  • Kubernetes: Dashboards giving insights into cluster health, deployments, and capacity usage
  1. Set up an administrator account to modify or add dashboards by signing in with username: admin and password: admin.
  • Before you expose the Grafana UI outside the cluster, make sure to change the password. You will be prompted to set a password on first login, and it can later be changed at http://localhost:3000/org/users.

Prometheus

Although Grafana provides a richer set of graphing functionality, Prometheus can still be useful for debugging.

  1. To open Prometheus, enter the following command:
  1. kubectl port-forward -n knative-monitoring \
  2. $(kubectl get pods -n knative-monitoring \
  3. --selector=app=prometheus --output=jsonpath="{.items[0].metadata.name}") \
  4. 9090
  • This starts a local proxy of Prometheus on port 9090. For security reasons, the Prometheus UI is exposed only within the cluster.
  1. Navigate to the Prometheus UI at http://localhost:9090

Metrics Troubleshooting

You can use the Prometheus web UI to troubleshoot publishing and service discovery issues for metrics.

  • To see the targets that are being scraped, go to Status -> Targets
  • To see what Prometheus service discovery is picking up vs. dropping, go to Status -> Service Discovery
  • To see a specific metric you can search for in the in search box on the top of the page