instant-serverless-pipeline

This repo is now archived and unmaintained

For alternatives see:

• https://github.com/symphoniacloud/github-codepipeline
• https://github.com/symphoniacloud/coffee-store-v2

Previous content:

Create a CodePipeline for a serverless app with one hand tied behind your back.

One of the chief benefits of Serverless apps is how quickly you can start a new application. That is, how long does it take to go from idea to production? And how long does it take to do that in a way that isn't a complete hack? We think you can do it very quickly. Without having to cut corners.

For a long time, we at Symphonia have believed that for any 'real' engineering project you have to start with some basic engineering building blocks:

• Source Control
• Code
• Tests
• Dependency management
• Build
• Packaging
• Deployment
• Continuous Integration
• Automated Deployment (Continuous Delivery)

This small project gets you started with all of the above on AWS, for a Serverless app, so that you can start concentrating on your code, as follows:

• For Source Control we stub with S3 for demo / training purposes, but will soon provide a Github version for more realistic workflows
• For code we give a starter lambda function in lib\index.js. This function makes a call to a remote resource, and returns a value based on that call.
• An example unit test is included at test\indexTest.js
• Dependency management is provided by NPM in package.json
• For build, also use the NPM script. See the tasks in the manual-deploy.sh file
• Packaging is a matter of simple zipping, defined in package.json .
• Deployment is via the Serverless Application Model (SAM) and CloudFormation, as defined in the sam.yml file
• Continuous Integration is supplied by CodePipeline and CodeBuild, defined in the deployment-pipeline/s3-source-pipeline.yml file or TODO-Github-version, triggered by an update to S3 or TODO-Github, and subsequent use of deployment-pipeline/buildspec.yml
• Automated Deployment is also supplied by CodePipeline, integrated with SAM

Using the skeleton

In your local environment make sure you have the AWS CLI setup, with a user with enough privileges to create CloudFormation stacks.

Using S3 as 'source control'

1. Clone this repo, change directory to deployment-pipeline, and run create-s3-source-pipeline.sh .
2. Find the instant-serverless-pipeline stack in CloudFormation and wait for it to be ready.
3. Change directory back to the project root and run commit-code-s3.sh
4. Find your pipeline in CodePipeline, and wait for it to start (triggered by S3, might take a minute) and complete.
5. Run deployment-pipeline/pipeline-test.sh to verify your application
6. Start programming!

Deploying manually

If you want to deploy the application manually, without CodePipeline, use the manual-deploy.sh script, specifying an S3 bucket to use as your CloudFormation package store. You can verify the application by running manual-test.sh .

Teardown

⚠️ These steps must be completed in order, or resources might be stranded.

1. Delete the application stack:

   $ aws cloudformation delete-stack --stack-name instant-serverless-app

2. Find the physical resource IDs for the build pipeline's S3 buckets

   $ aws cloudformation list-stack-resources --stack-name instant-serverless-pipeline \
         --query 'StackResourceSummaries[?ResourceType==`AWS::S3::Bucket`].PhysicalResourceId' \
         --output text

3. Delete those buckets using the the S3 web console (https://console.aws.amazon.com/s3/home)

4. Delete the build pipeline stack:

   $ aws cloudformation delete-stack --stack-name instant-serverless-pipeline

TODO

• Github version
• Tighten up privs in pipeline definition