workshop

Task 1

Terraform + AWS + Ansible

• create a new VPC with CIDR 172.16.0.0
• create two new subnets with CIDR 172.16.1.0/24 and 172.16.2.0/24 in two different availability zones
• create 5 new ec2 instances based on Ubuntu 18.04 (bionic)
• deploy the following Java application on these instances
• create a load balancer for the Java application on port 80
• setup route53 to host CNAME record of elb url

---

Task 2

Nginx + docker-compose + python

• creates a new Docker container running nginx and proxies all requests on /<container_name> to the appropriate container and port
  • there is only one nginx container running at all times
  • if the nginx container is down, it needs to be started
  • when a new application container is created the nginx configuration is updated to proxy requests to the new container
• creates a new Docker container running Java and deploys demo-0.0.1-SNAPTSHOT.jar from the previous step
  • the Docker container publishes container 8080 on a free port between 8000 and 8200
  • the container name is a unique identifier
  • container is only created if it does not exist
  • there can be multiple Docker containers with different names running at the same time

---

Task 3

github workflow + kustomize + kubectl: deployment + svc

• Implement a piece of software exposing a JSON document:

{
"id": "1",
"message": "Hello world"
}

when visited with a HTTP client

• Dockerize the application
• Put the application to kind's Kubernetes
• Create a second application, that utilizes the first and displays reversed message text
• Deployment docker image with github workflow and login github registry with k8s secret config.
• Update application with kubctl in script

---

Task 4

Python CLI

• Please consider the work you submit here a small, but production-ready deliverable, in the sense that you are happy to ship such code, tests and documentation.
• Write a Roman numeral converter that converts integer numbers into Roman numerals:

func(36)

Output: "XXXVI"

---

Task 5

Go tutorial on jupyter notebook

Create html from ipynb

jupyter nbconvert --execute task05/go_tuturial.ipynb --to html

html preview

---

Task 6

6.1 Creare a ruby script

It shows top 10 process in swap and sorts output by KB. It will ignore Zero values

6.2 Create a bach script

performs occasional test queries to the hostname "google.com" to each of the DNS servers configured in /etc/resolv.conf

6.3 Fix a client certificateds issue in nginx configure.

---

Task 7

setup backend api to handle POST/PUT/PATCH and query value metadata by filtering metadata key

• Technical scopes

github action
docker-compose
pytest
fastapi
minikube
make
Using helm to provision mongodb

---

Task 8

Istio 1.6.8 present canary deploy on minikube

---

Task 9

CI/CD within helm bluegreen deploy and release tag on gitlab pipelines

Gitlab pipeline

---

Task 10

build up a API to return GET request, and create a docker image on minikube, then create manifests to bring up the service and deployment.

---

Task 11

This program is going to be provided ​json​ lines as input in the ​stdin​, and should provide a ​json​ line output for each one — imagine this as a stream of events arriving at the authorizer.

$ cat operations
{"account": {"active-card": true, "available-limit": 100}} {"transaction": {"merchant": "Burger King", "amount": 20, "time": "2019-02-13T10:00:00.000Z"}}
{"transaction": {"merchant": "Habbib's", "amount": 90, "time": "2019-02-13T11:00:00.000Z"}}
$ authorize < operations
{"account": {"active-card": true, "available-limit": 100}, "violations": []} {"account": {"active-card": true, "available-limit": 80}, "violations": []} {"account": {"active-card": true, "available-limit": 80}, "violations": ["insufficient-limit"]}

---

Task 12

Provision kubeflow on minikube

---

Task 13

Use FastAPI and sqlalchemy to create shortened URL similar to https://goo.gl/

---

Task 14

Using terraform to create a private s3 bucket and a authorized user(IAM) to upload files. Using WhitelistIPs to grant user's exteral public IP address permission to access bucket.

⚠️ When you provision this task, you can not depend on STS token. Because there is a lack support of STS to create a new IAM user. ⚠️

# sample for whitelist to access s3 bucket
whitelistIPs = ["127.0.0.1/32"]

# get access_key from ssm
aws ssm get-parameter --name /system_user/backup-dev-uploader/access_key_id --with-decryption | jq .Parameter.Value

# get secret_key from ssm
aws ssm get-parameter --name /system_user/backup-dev-uploader/secret_access_key --with-decryption | jq .Parameter.Value

# terraform output
bucket_domain_name = "backup-dev-upload-task14.s3.amazonaws.com"

---

Task 15

You will find two applications: A Golang-based and a Java-based application. Both need to be containerized according to industry best practices. The Golang application needs to be compiled from source, while the Java application is delivered as a pre-built Jar file, runnable using Java 11. Both are providing an HTTP service, binding to all interfaces on port 8080, with the same endpoints:

Route	Description
/	A static site. Should not appear in the final setup
/hotels	JSON object containing hotel search results
/health	Exposes the health status of the application
/ready	Readiness probe
/metrics	Exposes Prometheus metrics of the application

Your challenge will be to provide a load balancer setup like the following:

                            +------------------>Java app
                            |
                            |
                            |
                            30% of traffic
                            |
                            |
User +-------> load balancer+
                            |
                            |
                            70% of traffic
                            |
                            |
                            |
                            +------------------>Go app

The traffic distribution should be as follows: 70% of the requests are going to the application written in Golang, 30% of the requests are going to the application written in Java. Also, each HTTP response needs to carry a custom header, called x-trv-heritage which indicates the application that responded. Your implementation must be runnable on a machine using x86_64 CPU architecture and must be built on top of Kubernetes. One should be able to at least see that the traffic distribution works as expected in some form. As a bonus, you can try to show other metrics, like CPU usage, memory utilization, and latency as well to compare the two services. Your implementation should:

• Build both container images locally
• Find a solution to make them available to a Kubernetes cluster
• Do not push them to a public registry on the internet!
• Setup an ingress solution of your choice
• Deploy both workloads
• Wait for the readiness of the system
• Run 100 requests against / of the applications under test

---

Task 16

You are tasked with the creation of a small infrastructure stack on
AWS:
* Deploy a redundant and scalable EKS cluster
* Deploy on the cluster a simple Web Server application, exposing on
the public internet a simple home page with a custom message.
* Provide basic monitoring for your infrastructure *[Optional Task]*
* Increase the scalability of the stack *[Optional Task]*
* Provide cost estimations *[Optional Task]*

Functional

EKS

• EKS cluster must consist of at least 3 Worker Nodes
• Worker Nodes should be distributed on at least 2 AZ
• Worker Nodes should be assigned to at least 2 Worker Groups
• You must run the latest version of EKS-1.21

Web Server

• Deploy, using Helm, a web server of your choice on the above running cluster.
• The deployment should span the 2AZ where EKS nodes are spread, and have a minimal redundancy.
• Customise the web server in order to show a home page with custom message like "Hello bot, welcome to your simple web page"
• Expose securely the page in order to be reached from public internet

---

Task 17

• Present k8s with digitalocean provider, using resource digitalocean_kubernetes_cluster and digitalocean_kubernetes_node_pool.
• Deploy bitnami/nginx-ingress-controller helm chart by using resource helm_release.

---

Task 18

• A simple example to present how helmfile can depend on includeing vaules.yaml from different directory to share the common setting between landscape.