Bugs API

The repository contains the code and instructions for building an deploying the Bugs API.

You will build a Kubernetes cluster (for dev and/or prod environments) hosting:

• the Bugs API server, implemented with FastAPI
• the database to store bugs, implemented with Redis

Prerequisites

The Bugs API will run on AWS EKS, so you need an AWS account and the following installed and correctly configured in your local machine if working locally (without a full CI/CD pipeline):

1. aws-cli with iam authenticator, configured with a IAM user with full privileges
2. docker
3. kubectl
4. terraform (version 0.11)

Overview of the repository structure

Folder	Content
terraform	Scripts to generate dev and prod clusters
api-server	Code and Dockerfile to build the Bugs API server
kubernetes	Kubernetes deployments
ci-cd	Proposals of Jenkins pipelines for a better CI/CD implementation

1. Create AWS resources

You will create:

1. A docker registry (ECR) to host the custom docker image for the API server
2. A Kubernetes cluster (EKS) to host the deployment. Development and production environments require two different clusters.

1.1 ECR creation

Move to the right folder: cd terraform/ecr

Execute the script (inspired by this: terraform init then terraform apply

At this point you'll have a new ECR called bugs. The ARN for this repository will be saved as the terraform output variable ecr_arn. To retrieve it: terraform output ecr_arn

You need docker to login into it by following this guide

1.2 EKS cluster creation

Move to the right folder: cd terraform/eks-cluster-<env> where env identifies the dev or prod stage.

Execute the script (inspired by this: terraform init then terraform apply

Generate kubeconfig and configmap for adding worker nodes from terraform outputs:

terraform output kubeconfig > ./kubeconfig
terraform output config_map_aws_auth > ./config_map_aws_auth.yaml

Update your kubectl configuration and apply the configmap to make worker nodes join the cluster:

export KUBECONFIG=./kubeconfig
kubectl apply -f ./config_map_aws_auth.yaml

After a while you should see all the nodes correctly registered to the cluster and ready by checking with this command: kubectl get nodes

2. Build the BUGS API server

Move to the right folder: cd api-server/eks-cluster

Build the custom docker image: docker build -t <image_name> .

Tag and push the image to your ECR:

docker tag <image_id> <ECR_ARN>/<image_name>
docker push <ECR_ARN>/<image_name>

3. Deploy everything

Move to the right folder: cd kubernetes

Create a class storage for AWS EBS: kubernetes apply -f gp2-storage-class.yaml

Create a namespace: kubernetes apply -f namespace.yaml

Deploy Redis: kubernetes apply -f redis.yaml

Deploy the API server: kubernetes apply -f api-server.yaml. Before applying the manifest, you need to configure the API server repository name and deployment environment name.

Alternative option: a basic Helm chart is in the helm folder.

4. Access the API server

The external IP address of the api-server service (autogenerated by the AWS LoadBalancer) can be retrieved with this command: kubectl -n bugs get svc

The JSON API is at /api/bugs

A simple web page to get all bugs and create a new one is at /web/bugs

An automatic Swagger UI will be at /docs