Deploying the Grafana Stack using Helm Charts

A comprehensive guide on deploying the Grafana Stack components (Prometheus, Loki, Tempo, and Grafana) using Helm charts.
Tags:

In this section, I will guide you through the process of deploying the Grafana Stack components (Prometheus, Loki, Tempo, and Grafana) and the Nginx ingress controller on your local Kubernetes cluster using Helm charts.

Deploying the kube-prometheus-stack Chart

The kube-prometheus-stack chart from the Prometheus community repository includes Prometheus, Alertmanager, and Grafana, as well as the necessary exporters and service monitors for Kubernetes cluster monitoring.

  1. Create a namespace for the Grafana Stack components:
kubectl create namespace monitoring
  1. Install the kube-prometheus-stack Helm chart:
helm install prometheus prometheus-community/kube-prometheus-stack --namespace monitoring

This command deploys Prometheus, Alertmanager, Grafana, and related components in the monitoring namespace.

Deploying Loki

To deploy Loki, use the loki chart from the Grafana repository.

  1. Install the loki Helm chart:
helm install loki grafana/loki --namespace monitoring

This command deploys Loki in the monitoring namespace.

Deploying Tempo

To deploy Tempo, use the tempo chart from the Grafana repository.

  1. Install the tempo Helm chart:
helm install tempo grafana/tempo --namespace monitoring

This command deploys Tempo in the monitoring namespace.

Deploying the Nginx Ingress Controller

The Nginx ingress controller is required to expose and access Grafana outside the Kubernetes cluster.

  1. Create a namespace for the Nginx ingress controller:
kubectl create namespace ingress
  1. Install the ingress-nginx Helm chart:
helm install nginx-ingress ingress-nginx/ingress-nginx --namespace ingress

This command deploys the Nginx ingress controller in the ingress namespace.

Accessing Grafana

To access Grafana, you need to create an ingress resource that maps a domain name to the Grafana service.

  1. Create a file named grafana-ingress.yaml with the following content:
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: grafana
  namespace: monitoring
  annotations:
    kubernetes.io/ingress.class: nginx
spec:
  rules:
  - host: grafana.local
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: prometheus-grafana
            port:
              number: 80

Replace grafana.local with the desired domain name for accessing Grafana.

  1. Apply the ingress resource:
kubectl apply -f grafana-ingress.yaml
  1. Update your /etc/hosts file to map the domain name to your local Kubernetes cluster’s ingress IP address. Find the IP address by running:
kubectl get ingress -n monitoring grafana -o jsonpath='{.status.loadBalancer.ingress[0].ip}'

Add the following line to your hosts file:

<IP_ADDRESS> grafana.local

Replace <IP_ADDRESS> with the actual IP address and grafana.local with the domain name you used in the grafana-ingress.yaml file.

  1. Access Grafana by opening your web browser and navigating to the domain name specified in the ingress resource (e.g., http://grafana.local). The default username and password for Grafana are admin. On your first login, you will be prompted to change the password.

Configuring Grafana Data Sources

After accessing Grafana, you need to configure the data sources for Prometheus, Loki, and Tempo.

  1. Configure the Prometheus data source by navigating to Configuration > Data Sources in the Grafana side menu, clicking on Add data source, and selecting Prometheus. Set the URL to http://prometheus-kube-prometheus-prometheus.monitoring:9090 and click on Save & Test.

  2. Configure the Loki data source by navigating to Configuration > Data Sources in the Grafana side menu, clicking on Add data source, and selecting Loki. Set the URL to http://loki.monitoring:3100 and click on Save & Test.

  3. Configure the Tempo data source by navigating to Configuration > Data Sources in the Grafana side menu, clicking on Add data source, and selecting Tempo. Set the URL to http://tempo.monitoring:3200 and click on Save & Test.

With the Grafana Stack components and the Nginx ingress controller deployed on your local Kubernetes cluster and Grafana configured with the appropriate data sources, you can now start exploring your cluster’s metrics, logs, and traces using Grafana dashboards.