Code Management

Cloud9:

Online IDE

Code Star:

Simplified Dashboard for all AWS pipeline tools

• Integration with JIRA, Github issues
• Supported Language: Java,HTML5, NodeJS, PHP, Ruby, Python, C#, Go
• Cloud9 for Web IDE
• Limited Customization
• Internally use Lambda, Beanstalk, EC2 to create environment

AWS CodeGuru

ML powered service for automate code review & performance recommend

• Reviewer: Automated code review using static code analysis
  • Identify Critical issue
  • Security holes
  • resource leaks
• Support
  • JAVA & Python.
  • Github, BitBucket & CodeCommit
• Profiler: Profile performance in real time in AWS or on premises
  • Optimize CPU usage
  • Space Optimization
  • Anomaly Detection

Code Commit:

Store code in git like repos

• Private Git Repo in AWS Cloud VPC
• No Repo Size Limit
• IAM Role to mange user access & used IAM Assume Role to share credential
• Encrypted at rest using KMS
• Encrypted in transit using SSH & HTTP
• Lambda & SNS can be integrated to trigger code Analysis for Sensitive Data

Feature	CodeCommit	GitHub
Authentication SSH & HTTPS	✅	✅
AWS CodeBuild Pipeline Support	✅	✅
Code Review(PR)	✅	✅
Security	IAM Role	Github Auth
Hosted by	AWS	Github
UI	Min	Complete

Code Pipeline

Visual Tool to Orchestrate Code COMMIT, BUILD, TEST, DEPLOY on AWS

• Source: Github, Bit Bucket, CodeCommit, S3, ECR
• Build: Jenkins, CodeBuild, TeamCity
• Test: CodeBuild, Device farm etc
• Deploy: CodeDeploy, Beanstalk, CLoudFormation, ECS, S3
• Staging: Production, Dev, Staging
• Monitoring: Using CLoudwatch Logs & Cloudtrail

ActionGroup:

Pipeline Stages can have multiple Action Groups

• Actions Can be defined parallel or Sequential for a stage

Code Artifact

hold build dependencies between pipeline stages in S3 Bucket

• Integration with Maven, Gradle, npm, pip etc
• Can proxy public Artifacts to Artifact repo

Code Build

Fully managed continuous integration service that compiles source code, runs tests, and produces software packages that are ready to deploy.

• With CodeBuild, you don’t need to provision, manage, and scale your own build servers, it is serverless.
• Monitor: Use Cloudwatch Metrics for build stats & Fail events
• Uses S3 to cache & store artifacts
• Supported environments:
  • Android, Java, .NET
  • Ruby, Python, NodeJS, PHP, Go
  • Docker to setup custom environment
• Run outside VPC but can be execute withing a VPC to access DB for testing. VPC Configuration need to be defined:
  • VPC ID
  • SubnetID
  • Security Group ID

CodeBuild Agent

run CodeBuild on local machine for troubleshooting

buildspec.yml

Build Instruction for code build at the root directory of project

|

env: env Variable

• plaintext
• parameter store: SSM parameter
• secret: AWS Secret Manger

phases

• install: warmup by install dependency
• pre_build: run before build
• build: run during build phase
• post_build: run after build to clean up

artifacts

• file to upload on S3

cache

• file to cache for next build

Code Deploy

Automate Provision & Deploy Code

CodeDeploy Agent

Continuously Poll CodeDeploy for Artifact to be deployed.

• EC2 Instance or on premise server must run code deploy agent
• Artifact to deploy must store in S3 or GitHub
• Agent Fetch Code and appsepc to deploy code
• Agent report success failure to CodeDeploy
• Compute Platform: Can deploy Code to
  • EC2/ On-premises:
  • Lambda
  • ECS.

appspec.yml

Instruction for code to deploy at the root directory of project

• files : how to copy from S3/ Github

  • source
  • destination

• hooks: instruction to follow on different phases

  • Application Stop
  • Download Bundle
  • Before Install
  • install
  • After Install
  • ApplicationStart
  • Validate Service <-- Run at the end to send Success/Fail Result
  • BeforeAllow Traffic
  • Allow Traffic
  • After Allow Traffic

Deployment Type

Method to deploy pp to Deployment Group

1. In Place

• Works With: EC2/ OnPremise

2. Blue/Green:

• Works with: EC2, Lambda, ECS
• Must use ELB

Terminology

• Application: revision to deploy

• Application Revision: Code + appspec.yml

• Target Revision: Most recent revision to deploy to Deploymnt Group

• Deployment Config: set of rules for success/failure

  • Configs
    • One at a time: stop if 1 EC2 fails
    • Half at a time: Take half down & upgrade
    • All at Once: Deploy at all result in downtime but fast
    • Custom: deploy to fix number of host75%

• Deployment Group: group of ec2 instance

• IAM Instance Profile: give EC2 access to GitHub & S3 to fetch code from

• Service Role: IAM role for codedeploy to perform action on EC2, ASG, ELB