Select a template to get a complete deployment walkthrough — pre-deploy checklist, step-by-step CLI and Console instructions, timing notes, and a post-deploy how-to for end users. Every guide below is written against the template's actual exported CloudFormation.
Build your architecture using drag-and-drop. Customize resource names and properties.
Click Export IaC → download CloudFormation JSON or YAML. Parameters make it renameable without breaking.
Follow this guide to deploy, then run each validation test in order. Timing notes tell you what to wait for.
S3 bucket served via CloudFront CDN. No servers. Globally distributed.
A clean network foundation: VPC with one public and one private subnet, Internet Gateway, route tables, S3 gateway endpoint, and flow logs. No NAT — nothing here has an hourly charge.
One t3.micro Amazon Linux 2023 instance in a public subnet, with SSM shell access, the CloudWatch agent, and daily AWS Backup snapshots. No load balancer, no database.
ALB → EC2 in VPC with RDS PostgreSQL. Classic 2-tier web application.
API Gateway (REST API) → Lambda → DynamoDB. Zero servers, scales to zero.
ALB → EC2 web tier → EC2 app tier → RDS PostgreSQL, each layer in its own subnet with its own security group.
ALB → ECS Fargate service (2 tasks) with execution + task IAM roles, plus RDS PostgreSQL.
CloudFront + S3 static frontend, plus a separate HTTP API (API Gateway → Lambda → DynamoDB) with Cognito auth.
SNS fan-out → SQS queue → Lambda processor → DynamoDB, with a dead-letter queue for failed messages.
Four CloudWatch billing alarms — $10 / $50 / $100 monthly plus a $5/day spike detector — publishing to one SNS email topic. Must be deployed in us-east-1.
Kinesis → Firehose → S3 data lake → Glue ETL → Redshift with Athena.
Multi-AZ ECS Fargate + Aurora Serverless v2 (PostgreSQL) + ElastiCache Redis, behind a WAF-protected internet-facing ALB.
Need to tear down a deployed stack?
Read the Decommissioning guide →