Monitor the EKS cluster Infrastructure

This is the GitHub dashboard and  the repository details are mentioned below,

https://raw.githubusercontent.com/kubernetes/dashboard/master/src/deploy/recommended/kubernetes-dashboard.yaml

You can run below command to configure the dashboard.

kubectl create -f https://raw.githubusercontent.com/kubernetes/dashboard/master/src/deploy/recommended/kubernetes-dashboard.yaml

After that you need to get the password for that run the

# kops get secrets kube --type secret -oplaintext

Then go into https://api.cloudwar.xyz/api URL and give the user name and the password. Here username is Admin and the password can be find by running the above command.

Select Kubeconfig and then select skip. You will be forwarded to the dashboard.

In this dashboard you can see the utilization and based on that you can decide whether you want to do the scaling up or down.

There are few other tools as well. you can try them also. But for me this dashboard covers my needs.

How to configure auto scale in EKS cluster

Exit from the master node and run the “kops edit ig --name=cloudwar.xyz nodes”

Change the MaxSize and the MinSize in to 3 instances then save the file. Then run

# kops update cluster cloudwar.xyz –yes

If you go into AWS EC2 console and check now you can see new instance was created

Error Creating the load balancer in EKS cluster

If you get an error like below that means you have a permission issue with the IAM role assigned.

To fix that issue you can first run “kubectl describe services hemalq-nginx” - hemalq-nginx is my the service and you will get actual issue with that.

Go into IAM and create a new policy

Select the JSON and paste the below mentioned policy

------------------------------------------------------------------------------------------------------------------

{

"Version": "2012-10-17",

"Statement": [

{

"Effect": "Allow",

"Action": "iam:CreateServiceLinkedRole",

"Resource": "arn:aws:iam::*:role/aws-service-role/SERVICE-NAME.amazonaws.com/SERVICE-LINKED-ROLE-NAME-PREFIX*",

"Condition": {

"StringLike": {

"iam:AWSServiceName": "SERVICE-NAME.amazonaws.com"

}

}

},

{

"Effect": "Allow",

"Action": [

"iam:AttachRolePolicy",

"iam:PutRolePolicy"

],

"Resource": "arn:aws:iam::*:role/aws-service-role/SERVICE-NAME.amazonaws.com/SERVICE-LINKED-ROLE-NAME-PREFIX*"

}

]

}

------------------------------------------------------------------------------------------------------------------

Then as we done in the initial stage, assigned this newly created poly into the Ubuntu instance and run the create load balancer command

Then once you go into the AWS Loadbalancers, you will see there is a new LB created. Under instances, you can find two instances also added.

Load Balancing the cluster

Here what I going to create an AWS load balance and point that to the created Pods. First run the below command

# kubectl expose deployment hemal-nginx --port=80 --type=LoadBalancer

Then run

# kubectl get services -o wide

If you get an error like below that means you have a permission issue with the IAM role assigned.

To fix this you need to add an IAM role with the required permission. The steps can be found here.

Deploy a Pod in to Kubernetes Cluster

First need to SSH into the Master node. The steps were covered in the previous post. Once you ssh into the master node you can run the below command. I have given the name as “hemal-nginx”

# kubectl run hemal-nginx --image=nginx --replicas=2 --port=80

To check the status of the pods you can run the below command

# kubectl get pods

Now the pods are deployed in to the AWS instance.

Net will configure the luster load balancer.

Configure EKS cluster

To configure the Kubernetes cluster, you can use the created Ubuntu server. steps are mentioned below,

1. SSH into your Ubuntu server

2. Go into Root

3. Update the kernel and install transport layer

# sudo apt-get update && apt-get install -y apt-transport-https

4. Add the required keys

# curl -s https://packages.cloud.google.com/apt/doc/apt-key.gpg | apt-key add -

# cat <<EOF>/etc/apt/sources.list.d/kubernetes.list

>deb http://apt.kubernetes.io/ kubernetes-xenial main

>EOF

5. Update the package

# apt-get update

6. Install Kubectl

# sudo apt-get install -y kubectl

7. Download Kops

# wget https://github.com/kubernetes/kops/releases/download/1.8.0/kops-linux-amd64

1. Add required permission to the Kops folder

# chmod +x kops-linux-amd64

# mv kops-linux-amd64 /usr/local/bin/kops

2. Put the public ssh key on home folder

# ssh-keygen

3. Create the cluster. This will not create the cluster but will create the require configuration which helps to create the cluster. You can change the DNS with your own one. And I’m creating this cluster on Souteastasia region and that details can be found under the instance description.

# kops create cluster --zones=ap-southeast-1a cloudwar.xyz

As you can see in the above print screen, before running the cluster create command you have to introduce the S3 bucket location where can store the cluster configuration details.

4. Set the S3 Bucket location

# export KOPS_STATE_STORE=s3://kuberneteshemalxyz.com

As mentioned this will create only the configuration files. Then you must run “kops update cluster cloudwar.xyz --yes” command to install the cluster.

• kops get cluster - This will list the configured cluster.
• kops edit cluster cloudwar.xyz – Can edit this cluster. Min max of the nodes
• kops edit ig --name=cloudwar.xyz nodes – Can edit the node instance group
• kops edit ig --name=cloudwar.xyz master-ap-s - Can edit master instance group
• kops update cluster cloudwar.xyz –yes - Configure the cluster

If you go to the AWS EC3 instance console now you can see New three VMs were automatically added there.

In Autos calling group there are two groups created.

You can validate the cluster by running “kops validate cluster”

The nodes details can be seen by running “kubectl get nodes --show-labels”

To ssh to the master Node you can run “ ssh -i ~/.ssh/id_rsa admin@api.cloudwar.xyz” This DNS need to be change with the DNS that you have configured.

Prerequisites for AKS cluster

DNS configuration on Route53

First need to configure the DNS configuration and if you already have a domain with you, then you can create a record set. But in my case, I have used my NS records in AWS but purchased from some other vendor. Then you can create the record sets. For Kubernetes cluster configuration it will automatically do the configuration and the permission

^{Create a S3 Bucket}

You need to create a S3 bucket and this will require to store the artefacts for Kubernetes cluster

Create an IAM role

This role will need to grant the permission to EC2 instance to create the instance and permission to view the S3 Bucket.

Go into IAM > Roles > EC2 > Administrator (Admin Permission) > Give a role name and save that.

Once this is done, you can add the created role and grant the required permission to the Ubuntu VM

Now we have configured the prerequisites to intel AKS cluster.