Deploying Prometheus and Grafana on a Test Kubernetes Cluster

This article demonstrates how to deploy a Prometheus monitoring stack and Grafana dashboard on a test Kubernetes (K8s) cluster.

First, verify the existing system pods:

[root@master ~]# kubectl get pods -n kube-system

Typical output shows core DNS, etcd, API server, controller manager, flannel, proxy, and scheduler pods all in Running state.

Create a directory for Prometheus data and switch to it:

[root@master ~]# mkdir /data/prometheus

[root@master ~]# cd /data/prometheus/

Deploy the node‑exporter daemonset and related services:

[root@master prometheus]# kubectl create -f node-exporter.yaml

Output:

daemonset.apps/node-exporter created

service/node-exporter created

Apply RBAC configuration, ConfigMap, and the Prometheus deployment and service:

[root@master prometheus]# kubectl create -f rbac-setup.yaml

clusterrole.rbac.authorization.k8s.io/prometheus created

serviceaccount/prometheus created

clusterrolebinding.rbac.authorization.k8s.io/prometheus created

[root@master prometheus]# kubectl create -f configmap.yaml

configmap/prometheus-config created

[root@master prometheus]# kubectl create -f prometheus.deploy.yml

deployment.apps/prometheus created

[root@master prometheus]# kubectl create -f prometheus.svc.yml

service/prometheus created

Check pod status again:

[root@master prometheus]# kubectl get pods -n kube-system

Most system pods remain Running , while the newly created node-exporter pods show ImagePullBackOff and the prometheus pod is in ContainerCreating state, indicating that the images are being pulled.

Deploy Grafana components:

[root@master prometheus]# kubectl create -f grafana-deploy.yaml

deployment.apps/grafana-core created

[root@master prometheus]# kubectl create -f grafana-svc.yaml

service/grafana created

[root@master prometheus]# kubectl create -f grafana-ing.yaml

ingress.extensions/grafana created

List services in the kube-system namespace to verify exposure:

[root@master prometheus]# kubectl get svc -n kube-system

Typical output shows Grafana exposed as a NodePort (port 3000), kube‑dns, node‑exporter, and Prometheus services with their respective cluster IPs and ports.

In Grafana, add a new Data Source of type Prometheus, configure the Prometheus server IP, import a ready‑made Prometheus dashboard template, and wait for metrics to appear.

After the dashboard loads, the monitoring platform is ready for use.

--- End of tutorial ---