DEV Community

Mario García
Mario García

Posted on • Updated on • Originally published at percona.com

Deploy Percona Monitoring and Management on Amazon EKS With eksctl and Helm

One of the installation methods we support for our database software is through Helm. We have a collection of Helm charts, in this repository, for the following Percona software:

  • Percona Operator for MySQL
  • Percona XtraDB Cluster
  • Percona Operator for MongoDB
  • Percona Server for MongoDB
  • Percona Operator for PostgreSQL
  • Percona Distribution for PostgreSQL
  • Percona Monitoring and Management (PMM)

Through this blog post, you will learn how to install Percona Monitoring and Management in a Kubernetes cluster on Amazon EKS using Helm and eksctl.

Requirements

For the installation of PMM with Helm, you will need:

  • Kubernetes 1.22+
  • Helm 3.2.0+
  • PV (Persistent Volume) provisioner support in the underlying infrastructure

If you want to install PMM in a Kubernetes cluster on Amazon EKS, you can use eksctl to create the required cluster. I published this blog post in the Percona Community blog where I explain how to use this tool.

For an easy way to deploy the Kubernetes cluster, check Percona My Database as a Service (MyDBaas), a Percona Labs project. I also recommend checking this article in our blog where the process of creating the cluster is described

MyDBaaS will help you with cluster creation. It will generate the configuration file needed for eksctl, or it can deploy the cluster to AWS.

To use eksctl you must:

  • Install kubectl
  • Create a user with minimum IAM policies
  • Create an access key ID and secret access key for the user previously created
  • Install AWS CLI and configure authentication (aws configure , from the command line)
  • Install eksctl

Create a Kubernetes cluster

To create the cluster, you will need to generate the configuration file for eksctl. Go to https://mydbaas.labs.percona.com/ and fill out the details of your cluster.

Figure 1: MyDBaaS - Kubernetes Cluster Configuration

  • Give your cluster a name. MyDbaas is the default value
  • Select the number of nodes. The value selected by default is 3 but you can create a cluster of up to 5 nodes
  • Write the instance type
  • Select the region. The default is us-west-2

If you don’t know what instance type to use, go to the Instance Type Selector and select:

  • Number of CPUs
  • Memory size
  • Region

Figure 2: MyDBaaS - Instance Type Selector

As stated on the website, this tool will only return the configuration file needed for eksctl. You can also provide your AWS credentials for the tool to deploy the EKS cluster.

After filling out the details, click on Submit and you will get the configuration file that will look like this:

addons:
- name: aws-ebs-csi-driver
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig
metadata:
  name: MyDbaas
  region: us-east-1
nodeGroups:
- desiredCapacity: 2
  iam:
    withAddonPolicies:
      ebs: true
      efs: true
      fsx: true
  instanceType: m5.large
  maxSize: 5
  minSize: 1
  name: ng-1
  preBootstrapCommands:
  - echo 'OPTIONS="--default-ulimit nofile=1048576:1048576"' >> /etc/sysconfig/docker
  - systemctl restart docker
  volumeSize: 100
  volumeType: gp2
Enter fullscreen mode Exit fullscreen mode

Then, create the cluster by running the following command:

eksctl create cluster -f cluster.yaml
Enter fullscreen mode Exit fullscreen mode

While running, eksctl will create the cluster and all the necessary resources. It will take a few minutes to complete.

Figure 3: eksctl Running

Cluster credentials can be found in ~/.kube/config . Try kubectl get nodes  to verify that this file is valid, as suggested by eksctl.

Install PMM with Helm

Once the cluster has been created and configured, you can install PMM using Helm.

  1. Add the repository
helm repo add percona https://percona.github.io/percona-helm-charts/
Enter fullscreen mode Exit fullscreen mode
  1. Install PMM
helm install pmm --set service.type="LoadBalancer" percona/pmm
Enter fullscreen mode Exit fullscreen mode

Once the PMM server runs, you must get its IP address. Run this command to get this value.

kubectl get services monitoring-service
Enter fullscreen mode Exit fullscreen mode

You will get an output similar to the following.

NAME                 TYPE           CLUSTER-IP      EXTERNAL-IP                                                              PORT(S)                      AGE
monitoring-service   LoadBalancer   10.100.17.121   a57d50410ca2f4c9d9b029da8f44f73f-254463297.us-east-1.elb.amazonaws.com   443:30252/TCP,80:31201/TCP   100s
Enter fullscreen mode Exit fullscreen mode

a57d50410ca2f4c9d9b029da8f44f73f-254463297.us-east-1.elb.amazonaws.com is the external IP and the one you need to access PMM from the browser.

Before accessing the dashboard of PMM, get the password.

export ADMIN_PASS=$(kubectl get secret pmm-secret --namespace default -o jsonpath='{.data.PMM_ADMIN_PASSWORD}' | base64 --decode)
echo $ADMIN_PASS
Enter fullscreen mode Exit fullscreen mode

The value of $ADMIN_PASS  is the password you need to log into the dashboard. The default user is admin.

Go to the browser and paste the external IP in the address bar.

Figure 4: PMM running on Kubernetes

Now you have PMM running in the cloud on Amazon EKS.

I would recommend that you check this article in our blog for more information about PMM in Kubernetes using Helm.

Conclusion

Through this blog post, you learned how to create a Kubernetes cluster with eksctl and deploy PMM using Helm, with the help of Percona MyDBaaS. The process would be the same for any of the Percona software in the collection of Helm charts.


This post was brought to you by Percona
Want to try Percona Monitoring and Management? Check our demo

Top comments (0)