Skip to content

Latest commit

 

History

History

exercise-15

README.md

15. Enable Metrics for Prometheus

See the changes in values yaml

There are two changes added:

 - "--metrics.prometheus=true"

Please note, the usage of Traefik Server and the service prometheus@internal

Create a service that expose metrics witin a cluster

kubectl apply -f traefik-dashboard-service.yaml

Validate metrics endpoint

kubectl get endpoints traefik-dashboard

Run temporary pod to get metrics.

kubectl run -it --rm --image alpine -- sh

curl traefik-dashboard:9000/metrics

15.1. Installing Prometheus stack using Helm

We will install Kube-Prometheus-Stack with all components: Prometheus, Alert Manager, Grafana and Metrics exporter.

Add the repository

helm repo add prometheus-community https://github.com/prometheus-community/helm-charts
helm repo update

Modify values YAML

  • Add extra chart that is needed for the example dashboard
  • Add the example dashboard as JSON. You can also import the chart manually once Grafana will be deployed.

Install Kube Prometheus Stack

helm install prometheus-stack  prometheus-community/kube-prometheus-stack -f promo-stack-values.yaml

Create Service Monitor that connects with Traefik Dashboard Service

kubectl apply -f traefik-service-monitor.yaml

Create port-forward to Promethues to explore data

kubectl port-forward service/prometheus-stack-kube-prom-prometheus 9090:9090

Go to http://localhost:9090/service-discovery to see whether Traefik target has been added.

Create port-forward to login to the Grafana and see the sample Traefik dashboard

kubectl port-forward service/prometheus-stack-grafana 8080:80

Login credentials can be taken from Kubernetes secrets:

kubectl get secrets prometheus-stack-grafana -o jsonpath='{.data.admin-user}'|base64 -d
kubectl get secrets prometheus-stack-grafana -o jsonpath='{.data.admin-password}'|base64 -d

Note

The following setup does not use persistence, which means once you restart the stack components all data will be lost. Review the original values file to enable data persistence for all components.

More information

See the blog post on Traefik's website: Capture Traefik Metrics for Apps on Kubernetes and Prometheus