Ship It on AWS: Cloud Deployment for Developers

09:00 – 10:15

See what AWS actually rents you

accounts and IAM least privilege, regions and availability zones, console vs CLI

Takeaway An account secured with least-privilege access and MFA, root credentials locked away, and the CLI talking to it.

10:30 – 11:45

Build the VPC that will otherwise block you

public and private subnets, route tables and internet gateway, security groups

Takeaway A network where the application tier is reachable and the database tier is not, proved from both sides.

12:00 – 13:00

Deploy to a bare EC2 by hand

launching and SSH, running the app under systemd, every manual step named

Takeaway Your application served from a raw virtual machine you set up yourself, and the written list of steps that day two automates away.

14:00 – 15:15

Put a load balancer and real HTTPS in front

target groups and health checks, ACM certificates, DNS

Takeaway A public HTTPS URL on a real domain with a certificate that renews itself, and a failing health check watched taking the instance out of rotation.

15:30 – 17:00

Give it a managed database, and test the backup

RDS in the private subnet, connection pooling, restores actually tested

Takeaway A managed PostgreSQL reachable only from the application, backed up and restored to a new instance, the restore verified rather than assumed.

09:00 – 10:15

Deploy the same app on Elastic Beanstalk

what a PaaS automates, environments and configuration, where the limits bite

Takeaway The same application live in minutes, with a written list of everything Beanstalk just did that you did by hand yesterday, and where you would outgrow it.

10:30 – 11:45

Run it as a container on ECS

ECR push and tagging, Fargate services behind the ALB, reading platform logs

Takeaway Your container from the Docker lab serving traffic behind yesterday's load balancer: the path you would actually take to work.

12:00 – 13:00

Keep secrets out of images, repos and logs

Secrets Manager and Parameter Store, rotation, the leaked layer

Takeaway A planted secret found in an old image layer, removed properly, and served from a secrets manager instead.

14:00 – 15:15

Ship on merge without an outage

GitHub Actions to ECR to ECS, rolling deploys and draining, rollback as a plan

Takeaway A merge taken to production by the pipeline under continuous load with no failed request, then rolled back in under a minute, on purpose.

15:30 – 17:00

Rebuild it all as code, then read the bill

Terraform plan and apply, destroy and rebuild, the three big line items

Takeaway The whole environment destroyed and rebuilt from the take-home Terraform repository, and a monthly cost estimate you can defend line by line.