Overview

This repository contains a lab for the mu workshop. The goal of this lab is to run a simple spring boot microservice on ECS.

Workspace setup

First, the following prerequisites should be installed on your machine:

• aws-cli: Install and configure
• git: Any git cli will do, however we will be using AWS CodeCommit for our source code. In order to use https URLs, you'll need to configure git to use the AWS CLI for generating temporary credentials:

git config --global credential.helper '!aws codecommit credential-helper $@'
git config --global credential.UseHttpPath true

• ssh: Any ssh client will do. You will need have a key to explore the instances that mu creates. The key needs to be imported into AWS via:

aws ec2 import-key-pair --key-name mu-key --public-key-material "$(cat ~/.ssh/id_rsa.pub)"

• mu: curl -s http://getmu.io/install.sh | sh

Pull down banana service

Let's start by pulling down the sample spring boot microservice:

git clone git@github.com:stelligent/mu-workshop-lab1.git banana-service

cd banana-service

Now, let's create a CodeCommit repo in our AWS account and push the service to the new CodeCommit repo. If you are working a region other than us-east-1, then be sure to update the URL for the CodeCommit repo:

aws codecommit create-repository --repository-name banana-service

git remote set-url origin https://git-codecommit.us-east-1.amazonaws.com/v1/repos/banana-service

Initialize mu.yml

We want to setup our mu.yml file so that mu can manage our service. The easiest way to start is with mu init. This doesn't actually do anything in your AWS account, just in your local mu.yml file. Pass the --env flag to have mu also setup the environment for your service:

mu init --env

This generates 2 new files, mu.yml and buildspec.yml. We need to make some updates to both to reflect run our microservice.

Configure environments

Use the Environments section of the wiki to configure the following for your environments:

• ECS container instance type: Configure as m4.large
• EC2 keypair: Configure as mu-key that we imported earlier

Configure service

Use the Services section of the wiki to configure the following for your service:

• Path pattern: update to only route /bananas*

Configure pipeline

Use the Pipelines section of the wiki to configure the following for your service pipeline:

• Build image: Update the image that CodeBuild uses to aws/codebuild/java:openjdk-8

Configure CodeBuild

Update the buildspec.yml to run gradle build.

Push changes to CodeCommit

git add --all
git commit -m "mu init"
git push

Create Pipeline

Now that our service is ready to deploy, let's create a pipeline for the service:

mu pipeline up

Explore and wait...

While we wait for the pipeline to run, let's explore what mu created:

• CodePipeline console - Open the AWS console and navigate to CodePipeline. Check out the different stages and actions in the pipeline.

  • What is the difference between the Artifact and Image actions in the Build stage?
  • What mu commands are being run in the Deploy actions?

• mu cli - Monitor the progress of the pipeline with mu

  • How can you view the logs from the CodeBuild executions hint?
  • How can you see the status of the pipeline with mu hint?

• CloudFormation console - Open the AWS console and navigate to CloudFormation. Check out the different stacks and resources created.

  • What is the difference between the pipeline, vpc, cluster and service stacks?
  • Is it possible to target an existing VPC rather than have mu create the VPC hint?
  • Check out the subnets in the VPC console. Are the ECS subnets public or private? ELB subnets?

• Service output - Use mu to view the logs for your service.

  • Do you see any errors in the service logs hint?
  • What do the errors mean?

• EC2 instances - Use the bastion host that mu created to login to your ECS instance

  • Get the bastion host and ECS instance IP via mu env show dev
  • Login to bastion host ssh -i ~/.ssh/id_rsa ec2-user@<bastion-ip>
  • Login to ECS instance
  • How many docker containers are running? (hint: docker ps)
  • How much CPU/memory are the containers using? (hint: docker stats)

• Test the service - Once the pipeline is successfully deployed to the dev environment, test it!

  • Get the base URL of the ELB via mu env show dev
  • Test the service via curl <baseurl>/bananas/
  • What does the response look like?