Skip to main content

​
📊 Monitoring Stack Setup with Prometheus, Loki, and Grafana on Kubernetes

This guide walks through the process of deploying a full monitoring stack (Prometheus, Loki, and Grafana) into a Kubernetes cluster using Helm.

​
🧾 Prerequisites

  • A running Kubernetes cluster
  • kubectl and helm installed and configured
  • Cluster access with sufficient privileges

​
🛠️ Step 1: Prepare Namespace and Repositories

Create a namespace for the monitoring stack and add the required Helm repositories: 
kubectl create namespace monitoring

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

​
🔍 Step 2: Explore Loki Chart and Customize Values

Search for Loki in the Helm repo: 
helm search repo loki
Export default values for customization: 
helm show values grafana/loki-stack > loki.yaml
NOTE: In loki.yaml, modify the grafana section: 
grafana:
  enabled: true
  image:
    tag: latest
This enables Grafana and optionally pulls the latest image tag.

​
🚀 Step 3: Deploy Loki Stack with Helm

Install or upgrade the Loki stack using the customized values: 
helm upgrade --install --values loki.yaml loki grafana/loki-stack -n monitoring --create-namespace
Loki UI will be available at the following internal URL: 
http://loki.monitoring.svc.cluster.local:3100

​
📈 Step 4: Install Prometheus

helm install prometheus prometheus-community/prometheus -n monitoring
Prometheus UI will be available at the following internal URL: 
http://prometheus-server.monitoring.svc.cluster.local

​
📊 Step 5: Install Grafana

helm install grafana grafana/grafana -n monitoring
Retrieve Grafana admin password: 
kubectl get secret --namespace monitoring grafana -o jsonpath="{.data.admin-password}" | base64 --decode ; echo
Expose Grafana locally via port-forward: 
kubectl port-forward service/grafana --namespace monitoring 3000:80
Alternatively, expose via NodePort (not recommended for production): 
kubectl patch svc grafana -n monitoring -p '{"spec": {"type": "NodePort"}}'
Access Grafana at http://<NodeIP>:<NodePort>

​
📉 Recommended Grafana Dashboards

You can import the following dashboards from Grafana Labs using their Dashboard IDs:
Dashboard NameDashboard ID
Kubernetes Addons (Prometheus)19105
Kubernetes Cluster (Global)15757
Kubernetes Namespaces View15758
Kubernetes Nodes View15759
Kubernetes Pods View15760
NVIDIA DCGM Exporter Dashboard12239
Kubernetes Nodes8171
Kubernetes Pod Metrics747
Kubernetes App Metrics1471

​
Argo Pipelines :

​
prometheus

project: default
source:
  repoURL: https://prometheus-community.github.io/helm-charts
  targetRevision: 58.5.0
  helm:
    parameters:
      - name: grafana.enabled
        value: 'false'
    releaseName: prometheus
    values: |
      grafana.enabled: false
  chart: kube-prometheus-stack
destination:
  server: https://kubernetes.default.svc
  namespace: monitoring
syncPolicy:
  syncOptions:
    - CreateNamespace=true