The Secret Life of AWS: The Assembly Line (AWS CodePipeline)

#aws#codepipeline#cicd#devops

Stop dragging zip files. Build a robot that deploys for you.

Part 38 of The Secret Life of AWS

Timothy hated Tuesdays. Tuesday was Deployment Day.

He sat at his desk, his palms sweating. On his screen was a folder containing three files:

• checkout-service-v2.zip
• checkout-service-v2-FINAL.zip
• checkout-service-v2-FINAL-fixed.zip

He took a deep breath.

1. Open AWS Console.
2. Navigate to Lambda.
3. Click "Upload from .zip file".
4. Select the file.
5. Click Save.
6. Pray.

He hit Save. The spinner spun.
Error: SyntaxError: Unexpected token

"No!" Timothy shouted. "It worked on my machine!"

Margaret appeared at his door, holding her tea. "Timothy, why are you shouting at the console?"

"I uploaded the wrong version," Timothy said, burying his face in his hands. "I must have clicked the old zip file. It works locally, I swear."

"You are hand-crafting your deployments," Margaret observed.

"I'm being careful!" Timothy argued.

"Humans are wonderful at creative thinking," Margaret said gently. "But for repetitive tasks? Sometimes we get tired and distracted. Robots never do."

"We need to stop carrying the code to the server by hand," she said. "We need an Assembly Line."

The Source (The Trigger)

Margaret navigated to AWS CodePipeline.

"A pipeline is a sequence of steps," she explained. "It starts with a trigger."

She pointed to Timothy's Git repository (GitHub).

"Right now, your code lives on your laptop. That is dangerous. The 'Source of Truth' should be the repository."

She created a new Pipeline: Checkout-Service-Pipeline.
Source Stage: GitHub (main branch).

"Now," she said, "The moment you push code to GitHub, the Assembly Line starts moving. You don't have to touch the AWS Console."

The Build (The Packaging)

"But my code needs to be packaged," Timothy said. "I have to run npm install to get the dependencies."

"Exactly," Margaret said. "And the pipeline should do that in a 'Clean Room,' not on your cluttered laptop."

She added a Build Stage using AWS CodeBuild.
She gave it a simple instruction file (buildspec.yml):

phases:
  install:
    commands:
      - npm install
  build:
    commands:
      - npm test
artifacts:
  files:
    - '**/*'

"This does two things," she explained.

1. It Tests: If npm test fails, the pipeline stops. The bad code never leaves the factory.
2. It Packages: It wraps up the clean code into an Artifact—a sealed box ready for delivery.

"Because this runs in a fresh environment every time," Margaret added, "'It works on my machine' is no longer a valid excuse."

The Delivery (The Robot)

"Now, we deliver the package," Margaret said.

She added the final Deploy Stage.
Deploy Provider: AWS Lambda.
Input Artifact: BuildArtifact.
Function Name: CheckoutFunction.

"Think of the Pipeline as a conveyor belt," Margaret said. "You put the raw material (code) on one end. The belt tests it, packages it, and installs it in the factory (Production). You never touch the factory floor."

The New Tuesday

The next Tuesday, Timothy finished his code. He didn't zip anything. He didn't open the AWS Console.

Without starting up AWS, he opened a local terminal and simply typed:
git push origin main

He watched his screen.

• Source: Succeeded (Just now).
• Build: In Progress... Succeeded.
• Deploy: In Progress... Succeeded.

Timothy checked his website. The new feature was there.
He hadn't dragged a single file. He hadn't held his breath.

"It feels... anti-climactic," Timothy said.

"Good engineering should be boring," Margaret replied. "You are no longer a mover, Timothy. You are the architect of the movement."

Timothy looked at the green checkmarks.
"But Margaret," he asked, "what if I want to see the website before it goes to customers? Like a practice run?"

"Ah," Margaret smiled. "Then you need a Staging Environment. But that is a lesson for another day."

---

Key Concepts

• AWS CodePipeline: A fully managed continuous delivery service that automates your release pipelines.
• CI/CD (Continuous Integration/Continuous Delivery): The practice of automating the build, test, and deploy phases.
• Artifact: The packaged output of the build process (e.g., a zip file) that is passed from one stage of the pipeline to the next.
• Build Stage (AWS CodeBuild): A managed build service that compiles source code, runs tests, and produces software packages.
• The "It Works on My Machine" Problem: Environmental drift caused by manual deployments. Pipelines solve this by building in a clean environment every time.

---

Aaron Rose is a software engineer and technology writer at tech-reader.blog and the author of Think Like a Genius.