All posts
Cloud · 6 min read · Apr 2, 2026

The AWS mistake that took down a demo environment

AWS infrastructure lesson

One morning, a demo environment became unavailable shortly after an EC2 instance was stopped and started. The application itself was healthy, Apache was running, Liferay was deployed, and logs showed no obvious errors. Yet users could not access the portal.

The issue was eventually traced to a simple AWS behavior. The instance had been using a dynamically assigned public IP address. When the server was restarted, AWS assigned a different public IP. DNS records still pointed to the previous address.

At first glance, the symptoms suggested an application problem. Teams investigated Apache configurations, SSL certificates, proxy settings, and application logs before realizing the root cause existed at the infrastructure layer.

The experience reinforced an important lesson: application availability depends on infrastructure assumptions that are often invisible until they fail. In our case, the portal, database, and web server were functioning correctly. The problem was that traffic was being directed to the wrong location.

The solution was straightforward. We associated an Elastic IP with the instance and updated deployment documentation to ensure DNS dependencies were explicitly reviewed during infrastructure changes.

bash
# Allocate a static Elastic IP and bind it to the instance
aws ec2 allocate-address --domain vpc

aws ec2 associate-address \
  --instance-id i-0a1b2c3d4e5f \
  --allocation-id eipalloc-0abc123

# Verify the public IP no longer changes across restarts
aws ec2 describe-instances \
  --instance-id i-0a1b2c3d4e5f \
  --query 'Reservations[].Instances[].PublicIpAddress'
Associating an Elastic IP gives the instance a stable public address across restarts.

This incident also highlighted the importance of operational checklists. A simple verification of public IP assignments after instance restarts could have prevented unnecessary troubleshooting.

Cloud platforms remove many operational burdens, but they also introduce behaviors that teams must understand. Sometimes the most disruptive outages are caused not by application bugs but by overlooked infrastructure details.

Contact

Let’s connect

Have a Liferay or Backend project you’d like to talk through? Fill in the form and I’ll get back to you.

Book Calendar