Chapter 20: AWS Lambda

AWS Lambda, the heart and soul of serverless computing on AWS.

If you’ve been following our journey so far (EC2 → Auto Scaling → Load Balancing → Messaging → Serverless intro), Lambda is where the magic really happens: you stop renting and managing servers forever and just write code that runs when something interesting occurs.

Let me explain it like we’re sitting together with a notebook and a second cup of chai — slow, detailed, full of real analogies, Hyderabad/India examples, how it works in 2026, pricing reality, common patterns, limitations, and a simple hands-on you can do right now.

1. What is AWS Lambda? (Simple + Official Definition – 2026)

AWS Lambda is a serverless compute service that lets you run code without provisioning or managing servers.

• You upload your code (called a function)
• You tell Lambda when to run it (triggers / events)
• Lambda executes your code in an isolated environment
• You pay only for the compute time used (billed in milliseconds)
• When no events happen → zero cost and zero infrastructure

Official short line (from aws.amazon.com/lambda/): “AWS Lambda is a compute service that lets you run code without provisioning or managing servers. Lambda automatically scales from zero to thousands of requests per second.”

In plain Hyderabad language: Instead of keeping a laptop/computer (EC2 instance) switched on 24/7 waiting for work, you give AWS a small piece of code and say: “Run this only when someone uploads a photo, or when an order is placed, or every 5 minutes, or when someone hits my API.” AWS wakes up a tiny container just for that moment, runs your code, sends the result, then shuts everything down. You pay ₹0.000 something per second of actual running time.

2. How Lambda Actually Works (Step-by-Step – 2026 View)

1. You write a function
   • Supported languages: Node.js, Python, Java, Go, .NET, Ruby, custom runtimes (even Deno, Rust via layers)
   • Code + dependencies (packaged as ZIP or container image)
2. You configure triggers (events that invoke Lambda)
   • API Gateway (HTTP requests)
   • S3 (file upload/delete)
   • DynamoDB streams / EventBridge / SQS / SNS
   • CloudWatch Events / Scheduled (cron-like)
   • Alexa, IoT, Step Functions, etc. (200+ AWS sources)
3. Invocation happens
   • Event arrives → Lambda creates execution environment (if cold)
   • Your code runs with event data as input
   • Returns response or side-effects (write to DB, publish SNS, etc.)
4. Execution environment lifecycle
   • Cold start: First run or after idle → 100 ms – few seconds (depends on language/size)
   • Warm: Subsequent calls reuse environment → <100 ms
   • Provisioned Concurrency / SnapStart → eliminate cold starts (pay extra)
5. AWS cleans up → environment destroyed after idle → zero cost

3. Key Limits & Features (Important to Know in 2026)

4. Pricing in 2026 (ap-south-1 Mumbai – Very India-Friendly)

• Free tier (forever):
  • 1 million requests/month
  • 400,000 GB-seconds compute time/month
• Beyond free:
  • Requests: $0.20 per 1 million (~₹16–17 per million)
  • Duration: $0.00001667 per GB-second (~₹0.0014 per GB-second)
  • Example: 512 MB function runs 200 ms → 0.1024 GB-seconds → cost ≈ ₹0.00014 per invocation

Realistic bill example (small Hyderabad startup):

• 500,000 API calls/month
• Average 400 ms duration, 512 MB memory
• Compute: ~20,480 GB-seconds → ~₹28
• Requests: ~₹8
• Total: ₹40–100/month (after free tier often ₹0)

Compare to EC2 t3.micro 24/7 (~₹1,500–2,000/month) — serverless wins big on variable/low traffic.

5. Real Hyderabad Example: Serverless Image Thumbnail Generator

Your college project / startup photo-sharing app:

Goal: When user uploads photo to S3 → automatically create thumbnail + store it.

Serverless flow (no servers):

1. User uploads photo.jpg to S3 bucket “hyderabad-photos”.
2. S3 event notification → triggers Lambda function “CreateThumbnail”.
3. Lambda code (Python example):
   • Receives event with bucket/key
   • Downloads photo from S3
   • Uses Pillow library to resize to 200×200
   • Uploads thumbnail-200.jpg back to S3
   • (Optional) Publishes to SNS “thumbnail created” → send push notification
4. Done in ~300–800 ms
5. During viral moment (10,000 uploads/hour) → Lambda scales to thousands of concurrent executions automatically
6. At night → zero invocations → ₹0

Cost: Usually ₹50–300/month even with moderate usage — vs running EC2 instance 24/7.

6. Quick Hands-On (Free Tier – Do This in 10 Minutes!)

1. Go to Lambda console → Create function → “Author from scratch”
2. Name: “HelloHyderabad2026”
3. Runtime: Python 3.12 or Node.js 20
4. Code (Python example):

5. Create test event → empty JSON → Test → see output
6. (Bonus) Add API Gateway trigger → get public URL → open in browser!

Cost? ₹0 (inside free tier).

Summary Table – Lambda Cheat Sheet (2026)

Lambda is the “just write code and let AWS handle the rest” service — the foundation of almost every modern serverless app in Hyderabad 2026.

Got it? Want next:

• Deep dive on cold starts & how to kill them?
• Full serverless CRUD API (API Gateway + Lambda + DynamoDB)?
• Lambda vs Fargate vs EC2 comparison?

Tell me — next lesson is ready! 🚀⚡