Thursday, May 26, 2022

How to setup monitoring on Kubernetes Cluster using Prometheus and Grafana | Setup monitoring on EKS Cluster using Prometheus and Grafana

What is Prometheus?

  • Prometheus is an open source monitoring tool
  • Provides out-of-the-box monitoring capabilities for the Kubernetes container orchestration platform. It can monitor servers and databases as well.
  • Collects and stores metrics as time-series data, recording information with a timestamp 
  • It is based on pull and collects metrics from targets by scraping metrics HTTP endpoints.

What is Grafana?

  • Grafana is an open source visualization and analytics software. 
  • It allows you to query, visualize, alert on, and explore your metrics no matter where they are stored.

Why System monitoring is critical?

  • Alerting: can give early warning of issues or problems before they become serious, costly or irreversible.
  • Visibility: Monitoring provides real-time insights into the health and performance of your applications and infrastructure.
  • Capacity Planning: By collecting and analyzing data over time, you can make informed decisions about capacity planning, such as when to add or remove nodes, allocate more resources, or scale your applications.

Please Watch this steps in YouTube Channel:

Prometheus Architecture

Key components:

    1. Prometheus server - Processes and stores metrics data
    2. Alert Manager - Sends alerts to any systems/channels
    3. Grafana - Visualize scraped data in UI

Installation Method:

The are are many ways you can setup Prometheus and Grafana. You can install in following ways:

1. Create all configuration files of both Prometheus and Grafana and execute them in right order.

2. Prometheus Operator - to simplify and automate the configuration and management of the Prometheus monitoring stack running on a Kubernetes cluster

3. Helm chart (Recommended) - Using helm to install Prometheus Operator including Grafana

Why to use Helm?

Helm is a package manager for Kubernetes. Helm simplifies the installation of all components in one command. Install using Helm is recommended as you will not be missing any configuration steps and very efficient. 

Pre-requisites:

Implementation steps

We need to add the Helm Stable Charts for your local client. Execute the below command:

helm repo add stable https://charts.helm.sh/stable


# Add prometheus Helm repo
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts


helm search repo prometheus-community

Prometheus and grafana helm chart moved to kube prometheus stack

Create Prometheus namespace
kubectl create namespace prometheus

Install kube-prometheus-stack

Below is helm command to install kube-prometheus-stack. The helm repo kube-stack-prometheus (formerly prometheus-operator) comes with a grafana deployment embedded.

helm install stable prometheus-community/kube-prometheus-stack -n prometheus
Lets check if prometheus and grafana pods are running already

kubectl get pods -n prometheus


kubectl get svc -n prometheus


This confirms that prometheus and grafana have been installed successfully using Helm.

In order to make prometheus and grafana available outside the cluster, use LoadBalancer or NodePort instead of ClusterIP.

Edit Prometheus Service

kubectl edit svc stable-kube-prometheus-sta-prometheus -n prometheus

Edit Grafana Service

kubectl edit svc stable-grafana -n prometheus

Verify if service is changed to LoadBalancer and also to get the Load Balancer URL.

kubectl get svc -n prometheus


Access Grafana UI in the browser

Get the URL from the above screenshot and put in the browser


UserName: admin
Password: prom-operator

Create Dashboard in Grafana

In Grafana, we can create various kinds of dashboards as per our needs.

How to Create Kubernetes Monitoring Dashboard?

For creating a dashboard to monitor the cluster:


Click '+' button on left panel and select ‘Import’.

Enter 12740 dashboard id under Grafana.com Dashboard.

Click ‘Load’.

Select ‘Prometheus’ as the endpoint under prometheus data sources drop down.

Click ‘Import’.


This will show monitoring dashboard for all cluster nodes



How to Create Kubernetes Cluster Monitoring Dashboard?

For creating a dashboard to monitor the cluster:


Click '+' button on left panel and select ‘Import’.

Enter 3119 dashboard id under Grafana.com Dashboard.

Click ‘Load’.

Select ‘Prometheus’ as the endpoint under prometheus data sources drop down.

Click ‘Import’.

This will show monitoring dashboard for all cluster nodes









    Create POD Monitoring Dashboard

    For creating a dashboard to monitor the cluster:


    Click '+' button on left panel and select ‘Import’.

    Enter 6417 dashboard id under Grafana.com Dashboard.

    Click ‘Load’.

    Select ‘Prometheus’ as the endpoint under prometheus data sources drop down.

    Click ‘Import’.



    This will show monitoring dashboard for all cluster nodes.



    Cleanup EKS Cluster

    Use the below command to delete EKS cluster to avoid being charged by AWS.

    eksctl delete cluster --name demo-eks --region us-east-1

    or Login to AWS console --> AWS Cloud formation --> delete the stack manually.

    you can also delete the cluster under AWS console --> Elastic Kubernetes Service --> Clusters
    Click on Delete cluster

    7 comments:

    1. Haha what to say here. I think the best document so far. I am a SQL server DBA and just started learning all this. I don't think other blogs will help me this much...Thank you so much for your time 🙏 will explore who site of yours. Thank you once again for the wonderful document.

      ReplyDelete
    2. if i want to un install the prometheus from the cluster, can you please suggest me the process of doing that as well.

      ReplyDelete
    3. how to change the port number after the installation of the svc... if the port numbers are overlapping with other service port numbers. how to overcome this issue

      ReplyDelete
    4. While installing kube-prometheus-stack chart with helm if I pass --set prometheus.service.type=LoadBalancer --set grafana.service.type=LoadBalancer then will the Grafana and Prometheus service be loadbalancers ? Kindly confirm because I heard somewhere that the chart only supports ingress and does not support loadbalancer.

      ReplyDelete

    Top 10 DevOps Popular Tools | Popular DevOps Tools You Must Know In 2024 | Learn DevOps Tools in 2024

    Here are the top 10 DevOps Tools to focus on to put your DevOps learning on a  faster track  and kick start your career quickly as a success...