In early April, I spun up a simple 3-tier architecture on AWS using Terraform. It was just for testing, and like any good engineer, I ran terraform destroy immediately after. Confident that everything was cleaned up, I moved on.

Until AWS sent me an email forecasting a bill of over $600. 😳

⚙️ What I Deployed

Using Terraform, I provisioned a basic 3-tier app architecture that included:

EC2 instances for compute
An Aurora RDS cluster
An Application Load Balancer
A NAT Gateway
All the supporting networking components (VPC, subnets, route tables)

All of this was deployed in the us-west-2 (Oregon) region.

💣 The Mistake

After my testing, I ran:

terraform destroy

Terraform confirmed the destruction was successful. To be thorough, I opened the AWS Console and verified that everything was gone. Or so I thought.

✅ Problem: I was checking us-east-1, not us-west-2.

Since AWS Console defaults to the last-used region, I didn't realize I was looking at an empty region. Meanwhile, the Aurora DB cluster and NAT Gateway were still active in Oregon, silently incurring charges.

💸 The $600 Surprise

Weeks later, my AWS billing dashboard showed:

RDS: $236.24
EC2-Other (NAT Gateway, etc): $102.24
Elastic Load Balancing: $1.20
Forecasted total: $606.96

When I reached out to @AWSSupport, they helped confirm the issue: the resources I thought were destroyed were still running — in the wrong region.

🧹 The Fix

Once I switched to the correct region (us-west-2) in the console, I immediately deleted:

✅ Aurora DB cluster
✅ NAT Gateway
✅ Any leftover EBS volumes and RDS snapshots

I also verified everything using:

terraform state list

and made sure no resources were still tracked or orphaned.

🧠 Lessons Learned

Here’s what I took away from this (expensive) experience:

✅ Double-check your region — both in the AWS Console and your Terraform config.
✅ Don't rely solely on terraform destroy output — cross-verify in the correct region.
✅ Use AWS Cost Explorer (daily granularity) to spot spikes in usage early.
✅ NAT Gateways and RDS can be very expensive — especially when idle.
✅ Snapshots are not automatically deleted — clean them up manually when needed.

🙏 Shout-out to AWS Support

Thanks to @AWSSupport for helping clarify what happened and pointing me in the right direction. If you ever run into weird billing issues, don’t hesitate to reach out — they were very responsive.

🧩 Final Thoughts

Cloud infrastructure is incredibly powerful, but it's easy to forget how real the costs can get. If you're learning Terraform or managing AWS resources, take the time to:

Understand where you're deploying
Automate cleanup checks
Monitor your cost dashboard regularly

Hopefully, my $600 mistake saves you some trouble (and money).

🚨 I Thought I Destroyed My AWS Infrastructure — But a $600 Bill Told Me Otherwise