👈

🧪 HANDS-ON 2

Event-Driven Processing

(Amazon S3 → AWS Lambda)

This lab teaches you true serverless behavior: 👉 code runs automatically when something happens.

1️⃣ What You Are Building (Big Picture)

You are building an automatic system where:

Whenever a file is uploaded to S3, a Lambda function runs automatically.

No API call. No button click. No server running.

🏗 Architecture

User / Application
 ↓ (Upload file)
Amazon S3 Bucket
 ↓ (Event notification)
AWS Lambda
 ↓
Process / Validate / Transform

2️⃣ Why This Lab Is VERY Important

This lab teaches:

✔ Event-driven architecture ✔ Asynchronous serverless execution ✔ Automation without APIs ✔ Real production use cases ✔ How AWS services talk to each other

This pattern is used everywhere in AWS.

3️⃣ Real-World Use Cases (Very Practical)

You’ll see this in real systems:

  • Image resize after upload
  • PDF validation
  • Virus scanning
  • Log processing
  • Data ingestion pipelines
  • File format conversion

4️⃣ Services Used & WHY

🔹 Amazon S3

  • Stores files (objects)
  • Triggers events automatically

🔹 AWS Lambda

  • Runs code when event occurs
  • No polling, no server

🔹 IAM

  • Allows S3 to invoke Lambda
  • Allows Lambda to read/write S3

5️⃣ Key Concept: Event-Driven ≠ API-Driven

API-Driven

Client → API → Lambda

Event-Driven

Event happens → Lambda runs

No client waits for response.

📌 This is asynchronous processing.

6️⃣ Step-by-Step Flow (CRITICAL)

Let’s walk through the exact packet / event flow.

🟢 Step 1: File Upload to S3

A file is uploaded:

  • Via AWS Console
  • CLI
  • Application
  • Lambda / EC2 / user

Example:

s3://my-bucket/uploads/image.jpg

🟢 Step 2: S3 Event Is Generated

S3 detects:

  • ObjectCreated event

Examples:

  • PUT
  • POST
  • Multipart upload

S3 creates an event message.

🟢 Step 3: S3 Triggers Lambda

S3:

  • Sends event to Lambda
  • Lambda is invoked asynchronously

📌 No retries from client 📌 S3 → Lambda is internal AWS call

🟢 Step 4: Lambda Receives Event Object

Lambda receives:

  • Bucket name
  • Object key (file path)
  • File size
  • Event type
  • Timestamp

Lambda now knows:

“Which file was uploaded, and where.”

🟢 Step 5: Lambda Processes the File

Lambda may:

  • Read file from S3
  • Validate content
  • Resize image
  • Rename file
  • Store output in another bucket

🟢 Step 6: Lambda Finishes & Stops

  • No server remains running
  • No idle cost
  • Execution ends

7️⃣ Lambda Event Object (Conceptual)

Lambda does NOT receive raw file.

It receives metadata like:

  • Bucket name
  • Object key

Lambda then:

  • Uses AWS SDK
  • Fetches file from S3 if needed

📌 Important exam point.

8️⃣ IAM Permissions (VERY IMPORTANT)

This lab teaches service-to-service IAM.

Who needs permission to do what?

🔹 S3 → Lambda

  • S3 needs permission to invoke Lambda

🔹 Lambda → S3

  • Lambda needs permission to:

    • Read object
    • Write object

IAM Role Flow

Lambda
 ↓ assumes
IAM Role
 ↓ allows
S3 access

Without IAM: ❌ Event fails ❌ Lambda errors

9️⃣ Stateless Nature (Reinforced)

Each file upload:

  • Triggers a new Lambda execution
  • No memory of previous files

State (if needed) stored in:

  • S3
  • DynamoDB

🔟 Scaling Behavior (Automatic & Powerful)

If:

  • 1 file uploaded → 1 Lambda execution
  • 1000 files uploaded → 1000 Lambda executions

AWS: ✔ Automatically scales ✔ No config needed

📌 This is why serverless is powerful.

1️⃣1️⃣ Error Handling (Important Concept)

What if Lambda fails?

  • S3 does NOT retry forever
  • Lambda retries automatically (limited)
  • Errors go to CloudWatch Logs

Best practice:

  • Use DLQ (SQS or SNS)
  • Add idempotent logic

1️⃣2️⃣ Cost Model (Why This Is Cheap)

You pay for:

  • S3 storage
  • Lambda execution time

You do NOT pay for:

  • Idle compute
  • Polling services

Event-driven systems are: ✔ Highly cost-efficient ✔ Scalable

1️⃣3️⃣ Common Mistakes (VERY IMPORTANT)

❌ Expecting Lambda to get file directly ❌ Forgetting IAM permission ❌ Processing very large files (> memory) ❌ Infinite loops (Lambda writes back to same bucket)

📌 Avoid triggering Lambda from the same bucket path it writes to.

1️⃣4️⃣ AWS Exam Points (MUST REMEMBER)

  • S3 can trigger Lambda
  • Event-driven architecture
  • Asynchronous invocation
  • Lambda receives metadata, not file
  • IAM role required
  • Auto scaling built-in

1️⃣5️⃣ Mental Model (REMEMBER FOREVER)

S3 = Event source
Lambda = Worker

Whenever S3 says:

“Something happened”

Lambda replies:

“I’ll handle it.”

1️⃣6️⃣ Why This Lab Is Essential

This lab teaches: ✔ Automation ✔ Event-driven thinking ✔ Real cloud workflows ✔ No API required

Many AWS systems are built like this.