AWS Serverless in Practice

Serverless means you don't manage servers. You upload code; AWS runs it when a request arrives.

The classic serverless trio

Client → API Gateway → Lambda → DynamoDB

• API Gateway — receives the HTTPS request and forwards it.
• Lambda — runs your function for a few milliseconds.
• DynamoDB — a fully-managed NoSQL database.

A tiny Lambda in Python

import json, boto3, uuid, os
table = boto3.resource("dynamodb").Table(os.environ["TABLE"])

def handler(event, _ctx):
    body = json.loads(event["body"])
    note = {"id": str(uuid.uuid4()), "title": body["title"]}
    table.put_item(Item=note)
    return {"statusCode": 200, "body": json.dumps(note)}

What you pay

• Lambda: first 1M requests + 400,000 GB-seconds free every month.
• DynamoDB: 25 GB storage free.
• API Gateway: 1M requests free for 12 months.

A personal-scale API is essentially free.

Beginner tips

• Cold starts — the first request after idle time may take ~500ms. Fine for background jobs, sometimes noticeable for user-facing APIs. Provisioned Concurrency or SnapStart fixes it.
• DynamoDB is NOT SQL — plan access patterns before designing the table (single-table design).
• Deploy with AWS SAM or Serverless Framework, not clickops.

Real-world example

TechNotesHub itself could be rebuilt fully serverless: static frontend on S3+CloudFront, API on Lambda + API Gateway, articles in DynamoDB. Running cost at 1,000 daily users: under $5/month.