AWS CodeDeploy

Building and testing your code is only half the story. You also need to get it onto your servers, containers, or Lambda functions — reliably and with the ability to roll back if something goes wrong. AWS CodeDeploy is a fully managed deployment service that automates application deployments to Amazon EC2 instances, on-premises servers, Lambda functions, and Amazon ECS services.

---

Why CodeDeploy?

Manual deployments are error-prone and stressful. Common problems include:

• Forgetting to update a configuration file on one server
• Deploying to all servers at once, causing a full outage when a bug slips through
• Having no automated way to roll back to the previous version
• Inconsistent deployment procedures between team members

CodeDeploy solves these problems by providing a repeatable, automated deployment process with built-in health checks and rollback support.

---

Core Concepts

Application

A CodeDeploy application is a logical container that groups your deployment configuration. You create one application per workload.

Deployment Group

A deployment group defines the target environment — which EC2 instances, Lambda function, or ECS service should receive the deployment. For EC2, you can target instances by:

• Tags — e.g. all instances tagged Environment=Production
• Auto Scaling Group — deploy to all instances in a specific ASG

Deployment Configuration

The deployment configuration controls how the deployment rolls out:

Configuration	Behaviour
AllAtOnce	Deploy to all targets simultaneously
HalfAtATime	Deploy to 50 % of targets, then the remaining 50 %
OneAtATime	Deploy to one target at a time
Custom	Define your own percentage or count

For production workloads, OneAtATime or a custom rolling strategy minimises risk.

Revision

A revision is the version of your application you want to deploy. For EC2/on-premises, this is typically a ZIP or tarball in S3 or a GitHub commit. For Lambda, it is a new function version or alias. For ECS, it is a new task definition.

---

Compute Platforms

CodeDeploy supports three compute platforms, each with different deployment strategies:

EC2 / On-Premises

Deploy application code and configuration to a fleet of servers. The CodeDeploy Agent must be installed on each target instance — it handles pulling the revision from S3 or GitHub and executing the deployment lifecycle hooks.

AWS Lambda

Deploy a new version of a Lambda function and shift traffic from the old version to the new version. CodeDeploy manages the alias traffic shifting.

Amazon ECS

Deploy a new task definition revision and shift traffic from the old task set to the new task set behind a load balancer.

---

The AppSpec File

The AppSpec file (appspec.yml) is the heart of a CodeDeploy deployment. It tells CodeDeploy what to deploy and how.

EC2 / On-Premises AppSpec

version: 0.0
os: linux

files:
  - source: /
    destination: /var/www/myapp

permissions:
  - object: /var/www/myapp
    owner: www-data
    group: www-data
    mode: "755"

hooks:
  BeforeInstall:
    - location: scripts/stop-server.sh
      timeout: 60
  AfterInstall:
    - location: scripts/install-dependencies.sh
      timeout: 120
  ApplicationStart:
    - location: scripts/start-server.sh
      timeout: 60
  ValidateService:
    - location: scripts/health-check.sh
      timeout: 120

Lifecycle Hooks (EC2 / On-Premises)

CodeDeploy executes hooks in a fixed order during each deployment: