Docker Build, Continuous Integration, and Deployment to Kubernetes using Kubectl CLI for a Node JS application

A simple Node JS application with unit tests and coverage reports using mocha and istanbul.

This repo demonstrates the following features:

• Set up serverless CI, i.e. on Shippable-provided infrastructure
• Set up CD pipelines for a Kubernetes cluster running on Amazon EC2
• Perform CI tests
• Perform docker build and push image to Docker Hub
• Automatically deploy image to TEST environment
• Manually deploy image to PROD environment
• Set up runSh job type in Shippable using kubectl CLI (for declarative deploy job, see this sample project)

Run CI for this repo on Shippable

• Fork this repo into your source code account (e.g. GitHub)
• Create an account (or login) on Shippable with your SCM account
• Create an integration on Shippable for your Docker Hub account
• Update the CI configuration in shippable.yml file with your integration names (see comments in file)
• Follow these setup instructions to enable your forked repo for CI and run a build

Add Continuous Delivery pipelines to deploy to Kubernetes

• Create an integration for Kubernetes
• All pipeline config is in shippable.resources.yml and shippable.jobs.yml. Check these files and update config wherever the comment asks you to replace with your specific values
• Follow instructions to add your Continuous Deployment pipeline
• Right-click on the runSh job in the SPOG view named 'deploy_kubernetes_runcli_test' and run the job
  • This demo uses custom scripting jobs called 'runSh' jobs in Shippable - learn how moe about runSh jobs
• Your app should be deployed to your Kubernetes cluster as a Test pod
• Follow instructions to connect your Continuous Integration project to your Continuous Delivery pipelines(for this demo, just uncomment the trigger integration in shippable.yml)
• Right-click on the runSh job in the SPOG view named 'deploy_kubernetes_runcli_prod' and run the job to deploy to a Prod pod
• Make a change to your forked repo and commit to GitHub - watch your pipeline automatically execute CI with push to Docker Hub and automatic deployment to the Test environment in Kubernetes
• Then right-click to deploy the newest changes to the Prod environment

Your end-to-end pipeline is complete! Now, any change you make to the application will be deployed to your Kubernetes Test pod and be ready to manually deploy a Prod pod, as well.

CI console screenshot

Kubernetes integration screenshot

CD Pipeline SPOG screenshot

Launch a Kubernetes cluster for this demo

Instructions for launching a Kubernetes cluster using KOPS:
https://kubernetes.io/docs/getting-started-guides/kops/

Determine settings for your cluster:
YOUR_CLUSTER_DOMAIN_NAME="your.cluster-name.com"
YOUR_ZONE="your_aws_zone" # e.g. us-east-1c

Build the cluster configuration (replace the zone and cluster name with your values):

kops create cluster --zones=$YOUR_ZONE $YOUR_CLUSTER_DOMAIN_NAME 

Create the cluster in AWS:

kops update cluster $YOUR_CLUSTER_DOMAIN_NAME 

To add a Kube UI:

kubectl create -f https://raw.githubusercontent.com/kubernetes/kops/master/addons/kubernetes-dashboard/v1.4.0.yaml
kubectl proxy

Then navigate to http://127.0.0.1:8001/ui

Delete the cluster:

kops delete cluster $YOUR_CLUSTER_DOMAIN_NAME --yes

To add ability to pull images from private registry:
https://kubernetes.io/docs/user-guide/images/#specifying-imagepullsecrets-on-a-pod