Installing Prometheus in a Kubernetes Cluster

Prometheus is an open-source monitoring and alerting system widely used in cloud native environments. In this article, we will cover the steps to install Prometheus in a Kubernetes cluster.

Prerequisites

A running Kubernetes cluster kubectl installed on your machine and configured to communicate with your cluster Familiarity with the Kubernetes namespace and YAML files

Step 1: Create a Namespace

To install Prometheus, we need to create a new namespace in our Kubernetes cluster to keep the resources separate.

kubectl create namespace prometheus

Step 2: Install the Prometheus Operator

The Prometheus Operator is a tool that makes it easy to manage and operate Prometheus clusters on top of Kubernetes. To install the Prometheus Operator, run the following command:

kubectl apply -f https://raw.githubusercontent.com/prometheus-operator/prometheus-operator/main/bundle.yaml -n prometheus

Step 3: Create a Prometheus Deployment

Now that we have installed the Prometheus Operator, we can create a Prometheus deployment in our cluster. This is done by creating a Prometheus custom resource definition (CRD).

Here is an example Prometheus CRD:

apiVersion: monitoring.coreos.com/v1
kind: Prometheus
metadata:
  name: example
  namespace: prometheus
spec:
  version: v2.22.1
  serviceAccountName: example-prometheus
  ruleSelector:
    matchLabels:
      prometheus: example
  resources:
    requests:
      memory: 400Mi
  replicas: 2
  retention: 30d
  serviceMonitorSelector:
    matchLabels:
      app: example

Apply the CRD using the following command:

kubectl apply -f prometheus.yaml -n prometheus

Step 4: Verify the Installation Once the Prometheus deployment is created, we can verify the installation by checking the pods in the prometheus namespace.

kubectl get pods -n prometheus

You should see two Prometheus pods in the Running state.

Step 5: Access the Prometheus Web UI To access the Prometheus web UI, we need to expose the Prometheus service using a Kubernetes LoadBalancer. Here is an example service definition:

apiVersion: v1
kind: Service
metadata:
  name: example-prometheus
  namespace: prometheus
  labels:
    prometheus: example
spec:
  selector:
    prometheus: example
  ports:
    - name: web
      port: 9090
      targetPort: web
  type: LoadBalancer

Apply the service definition using the following command:

kubectl apply -f prometheus-service.yaml -n prometheus

Once the service is created, you can access the Prometheus web UI by using the LoadBalancer IP address and port 9090.

In this article, we covered the steps to install Prometheus in a Kubernetes cluster. With Prometheus, you can monitor your applications and infrastructure and receive alerts in case of any issues.