Udacity Project - Refactor Udagram Into Microservices

Refactoring udagram initial monolitic application into separate microservices as follows:

• Divide the application into smaller services and extend it with additional service and automatic continuous integration and continuous delivery.
• Containerize the application, create the Kubernetes resource and deploy it to a Kubernetes cluster.
• Extend the application with deployments and be able to do rolling-updates and rollbacks.

Usage

Installation

• Prerequisite: It is best to have Ubuntu 18.04, Docker 18.09, KubeOne v0.9.0, Terraform 0.12.8 installed before starting the project.
• Once the repository has been downloaded, you will need to first install all dependencies by running npm i. You will need to repeat this for all Frontend, Restapi-Feed and Restapi-user folders.
• Also, ensure all environment variables are set up in your ~/.profile.
• For Docker:
  • First, install Docker in your system.
  • Build images using docker-compose -f docker-compose-build.yaml build --parallel
  • Push the docker images to DockerHub using docker-compose -f docker-compose-build.yaml push.
  • Run application locally using docker-compose -f docker-compose.yaml up. Note you have to export enviroment varibales needed for the application first. Fill secrets.example.sh and Run source secrets.example.sh
  • Go to Localhost:8100 to test.
  • Docker deployment complete.
• For Kubernetes:
  • Install Kubernetes using the Kubeone instructions given here: https://github.com/kubermatic/kubeone/blob/master/docs/quickstart-aws.md
  • Follow the setup videos closely, and export the KubeConfig.
  • Check for the proper setup using kubectl get pods.
  • Set up the secret keys and configmaps for the project. These are the .yaml files in the k8s folder.
  • Make sure the .yaml files are also pointing to the right Docker images which you set up previously.
  • Use kubectl apply -f [file-name] on the secret key, config map files.
  • Use kubectl apply -f [file-name]' again for the remaining .yaml files. If all is set up correctly, you will see all pods running. Test using kubectl get rs, kubectl get deployment, and kubectl get pods`.
  • Forward the deployment port to Localhost:8080 using kubectl port-forward [name]/reverseproxy-[id] 8080:8080.
  • Go to Localhost:8080 to test.
  • Kubernetes deployment completed.
• For Travis:
  • Ensure your Github account is linked to Travis. This is done via signing up for a Travis account, then from within linking the two together.
  • Push your folders to Github repository. No need for node_modules nor secrets to be included in the repository.
  • Ensure the .travis.yml file is in the root folder of the Github repo.
  • Go to Travis, then search for the repository, and then set up new build.
  • Travis deployment completed.

Dependencies

There are dependencies on having an S3 bucket provisioned and an RDS postgres instance provisioned if you want to run this locally on your machine

Images for evidence

• Attached are the following images which demonstrate the functionality of the project:
  • Running application;
  • DockerHub images;
  • Kubernetes running pods; and
  • Travis deployment completed.