Advanced Deployment Strategies in AWS

This guide covers the critical strategies for deploying applications across various AWS compute platforms, focusing on the methodologies, tools, and best practices required for the DOP-C02 exam.

Learning Objectives

By the end of this module, you should be able to:

• Distinguish between mutable and immutable deployment patterns.
• Configure AWS CodeDeploy for EC2/On-premises, ECS, and Lambda platforms.
• Implement Blue/Green and Canary deployment strategies to minimize downtime.
• Define lifecycle event hooks in appspec.yml for different compute types.
• Select appropriate storage patterns (EBS, EFS, S3) based on application persistence needs.

Key Terms & Glossary

• AppSpec File: A configuration file (YAML/JSON) used by CodeDeploy to manage the deployment lifecycle.
• In-place Deployment: An update method where the application on each instance is stopped, updated, and restarted. (EC2 only).
• Blue/Green Deployment: A strategy where a new environment (Green) is provisioned alongside the old one (Blue). Traffic is shifted once the Green environment is validated.
• Canary Deployment: A phased deployment where a small percentage of traffic is shifted to the new version before a full rollout.
• Immutable Infrastructure: A pattern where servers are never modified after deployment; updates involve replacing the entire instance or container.
• Mutable Infrastructure: A pattern where existing servers are updated in-place (e.g., via SSH or configuration management).

The "Big Idea"

In a DevOps environment, the goal is Zero-Downtime Deployment. AWS provides a spectrum of tools to achieve this, ranging from the highly customizable CodeDeploy to the automated orchestration of ECS and Lambda. The choice of strategy depends on the Compute Platform (is it a server, a container, or code?) and the Risk Tolerance (how much traffic can we risk on unvalidated code?).

Formula / Concept Box

Hierarchical Outline

• I. Deployment Methodologies
  • A. Mutable (In-Place): Updates existing resources. Slower recovery, risk of "configuration drift."
  • B. Immutable (Replacement): Replaces resources. Faster rollbacks, consistent state.
• II. AWS CodeDeploy Architecture
  • A. Deployment Groups: Logical tags for environments (Staging vs. Production).
  • B. AppSpec Lifecycle Hooks:
    • EC2 Hooks: BeforeInstall, AfterInstall, ApplicationStart, ValidateService.
    • Lambda Hooks: BeforeAllowTraffic, AfterAllowTraffic.
• III. Platform-Specific Strategies
  • A. Amazon EC2: Supports Auto Scaling Group (ASG) integration for automatic deployment to new instances.
  • B. Amazon ECS: Utilizes Task Sets; requires an ELB to swap traffic between Blue and Green.
  • C. AWS Lambda: Uses Aliases and Weights to shift traffic (e.g., Canary10Percent5Minutes).

Visual Anchors

Blue/Green Deployment Workflow

Infrastructure Modification Patterns

Definition-Example Pairs

• Linear Deployment: Traffic is shifted in equal increments over time.
  • Example: Shifting 10% of Lambda traffic every 10 minutes until 100% is reached.
• AppSpec ValidateService Hook: A script used to verify the deployment was successful.
  • Example: A curl command sent to the local endpoint localhost:80/health to ensure the web server is responding before marking the deployment 'Succeeded'.
• EC2 Image Builder: An automated service to create and maintain "Golden Images" (AMIs).
  • Example: A pipeline that triggers every month to patch an Amazon Linux AMI, installs security agents, and shares it with the Production AWS account.

Worked Examples

Scenario: Converting an EC2 In-place Update to Blue/Green

The Problem: A company updates their EC2 instances using git pull on each server. This causes 5 minutes of downtime and occasionally leaves servers in inconsistent states.

The Solution:

1. Artifact Preparation: Use CodeBuild to bundle the application into a .zip file and push to S3.
2. AppSpec Update: Create an appspec.yml that defines the destination for files and uses the ApplicationStart hook to launch the service.
3. CodeDeploy Configuration: Create a Blue/Green deployment group. Select the Auto Scaling Group currently running the app.
4. Traffic Shifting: Configure CodeDeploy to use an Application Load Balancer (ALB).
5. Execution: When deploying, CodeDeploy clones the ASG, installs the new version on the new instances, and swaps the ALB target group once ValidateService passes.

Checkpoint Questions

1. Which compute platform allows for In-place deployments via AWS CodeDeploy?
2. In a Lambda deployment, which AppSpec hook would you use to run a functional test before the traffic starts shifting?
3. What is the primary difference between a Canary and a Linear deployment strategy?
4. True or False: The CodeDeploy agent must be installed on ECS Fargate tasks to perform a deployment.

▶Click to see Answers

1. Amazon EC2 / On-premises.
2. BeforeAllowTraffic.
3. Canary shifts a single chunk (e.g., 10%) then the rest; Linear shifts increments at set intervals (e.g., 10% every X minutes).
4. False. ECS deployments are handled via the ECS service and task sets; no agent is needed within the container.

Muddy Points & Cross-Refs

• AppSpec vs. BuildSpec: Remember that buildspec.yml is for CodeBuild (compiling code/creating artifacts), while appspec.yml is for CodeDeploy (placing files/running hooks).
• ECS Deployment Errors: A common failure point is the TaskDefinition execution role. Ensure the role has permissions to pull the image from ECR, otherwise, the Blue/Green swap will time out.
• Rollbacks: CodeDeploy rollbacks are automatic by default on failure, but for Blue/Green, you must specify whether to terminate the original instances immediately or keep them for manual inspection.

Comparison Tables

Deployment Method Comparison