2 Eks Basics

EKS Basics

Resources

Cost of EKS

Resource	Cost
Control plane	USD$73/month
Per node	Changes per node

Note: flat rate is even if there are no nodes.

The best way to calculate is to go to the AWS calculator.

Example pricing:

Resources	Calc	Numbers	Total
EKS with 3 m5.large in us-east-1	EKS Control Plane + Worker Nodes (m5.large X 3)	$73 + (3 * $81.76)	$318.28
EKS with 3 t4g.large in us-east-1	EKS Control Plane + Worker Nodes (t4g.large X 3)	$73 + (3 * $58.4)	$248.2

Ways to spin up a cluster

AWS Console
CloudFormation
AWS CLI
AWS CDK
eksctl CLI

eksctl

The best way to learn by getting hands-on quickly.

CLI tool for creating clusters with EKS
Abstracts lots of stuff - VPC, subnet, sec group etc using CloudFormation

Command	Brief Description
`eksctl create cluster`	Create EKS cluster with one nodeground containing 2 m5.large nodes
`eksctl create cluster --name <name> --version 1.15 --node-type t3.micro --nodes 2`	Creates 2 tw.micro nodes with Kubernetes 1.15
`eksctl create cluster ... --managed`	Created cluster with managed node group
`eksctl create cluster --fargate`	Create Cluster with Fargate Profile

Install + configure

# For Mac
$ brew install eksctl
# ... installs

kubectl

CLI for running commands against a cluster on k8s Resources
Communicate via cluster API server
Works for any k8s cluster - EKS, k8s on EC2, GKE etc.

Spinning up the first cluster

In the example, the following is run:

$ eksctl create cluster --name eksctl-test --nodegroup-name ng-default --node-type t3.micro --nodes 2
# ... response based on creating the cluster

It will take about 5-10 minutes and creates two CloudFormation templates.

There was an example yaml that eksctl can also use:

# cluster.yaml
apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig

metadata:
  name: basic-cluster
  region: eu-north-1

nodeGroups:
  - name: ng-1
    instanceType: m5.large
    desiredCapacity: 10
    volumeSize: 80
    ssh:
      allow: true # will use ~/.ssh/id_rsa.pub as the default ssh key
  - name: ng-2
    instanceType: m5.xlarge
    desiredCapacity: 2
    volumeSize: 100
    ssh:
      publicKeyPath: ~/.ssh/ec2_id_rsa.pub

This can be created using eksctl create cluster -f cluster.yaml.

To delete we can run eksctl delete cluster --name <NAME>.

Pod limit in a node

Max number of allow pods depends on EC2 instance type.
Bigger the instance, more pods.

You can see a list here

By default, you will see that the defaults already have some pods running.

Running kubectl get pods -n kube-system would show you the system pods required to run.

In an experiment, there were 10 pods made and noticeably there were no pods running.

Managed Nodegroups

This is an EKS-exclusive feature.

Say the follow happens:

k8 version needs to be updated
New security patch
EC2 AMI needs to be updated
- You need to create AMIs (nodes would drain pods and app will be down and you need to orchestrate HA)

Managed nodegroups automate the deployment of updates AMIs with security patches, CVEs.

No app downtime
No overhead of user management orchestratration
Autoscaling group done behind the scenes

How does it do it?

Worker Node k8s updated with one click/one API call
AWS provides AMI with security updates. No patching/rehydrate effort for you.

Demo

In the example, the following is run:

# 1.14 for used to update
$ eksctl create cluster --name eksctl-test --version 1.14 --nodegroup-name ng-default --node-type t3.micro --nodes 2 --managed
# ... response based on creating the cluster

After creating, the cluster control plane version was updated from the EKS dashboard.

After updated, there is a message on the dashboard letting you know that there is a new version for AMI groups.

You can head down and update that from there.

Helm

Helm solves the problem of multiple manifest files and packaging re-useable sets of charts.

$ helm install x stable/wordpress
# ... installs

Helm is a package manager for k8s
Packages are called Charts
Helm Charts help define, install and upgrade complex K8s apps
Charts can be versioned, shared and published
Helm charts can accept input params
- kubectl needs template engine to do this
Popular packages already available

The structure:

├── Chart.yaml
├── values.yaml
└── templates/
    ├── service.yaml
    └── deployment.yaml

Command	Description
`helm search repo <SEARCH>`	Search local metadata for chart
`helm search hub <SEARCH>`	Search greater hub for chart
`helm repo add <NAME> <URL>`	Add a chart to your metadata
`helm pull bitnami/nginx --untar=true`	Pull down nginx server code
`helm install <NAME> bitnami/nginx`	Install nginx chart under desired name