Difficulty: Easy
Description
We identified several EC2 instances in your account that are idle. Stopping idle EC2 instances can help reduce your EC2 bill.
Selection Criteria
The EC2 instance has a maximum CPU utilization of less than 2% over the last 31 days
The EC2 instance is not part of an ASG (Auto Scaling Group)
Expected Saving
Stopping an instance eliminates 100% of its compute cost. EBS volumes attached to the instance continue to be billed.
Example: A t3.medium instance running 24/7:
On-Demand rate: $0.0416/hr × 730 hrs/month = ~$30.37/month saved
EBS volumes remain billed: e.g. 20 GB gp3 = $1.60/month (still charged after stop)
Net savings: ~$28.77/month per instance
⚠️ Note — Elastic IP: An Elastic IP associated with a stopped instance incurs charges (~$0.005/hr = ~$3.65/month). Factor this into the net savings if applicable.
Operational Impact
Downtime: Yes — the instance will be stopped and all applications hosted on it will become unavailable.
Breaking changes: All running processes are interrupted; in-memory state (RAM) is lost. Data stored on EBS volumes is preserved.
Recovery: Restart the instance at any time from the AWS Console or CLI — EBS volumes and their data are fully intact.
⚠️ Warning — EBS volumes: Attached EBS volumes continue to be billed after the instance is stopped (~$0.08–$0.10/GB/month for gp2/gp3).
⚠️ Warning — Elastic IP: If an Elastic IP is associated with the stopped instance, it will incur hourly charges (~$0.005/hr). Consider releasing it if the instance is to remain stopped.
⚠️ Warning — Auto-restart: Verify that no scheduled tasks (cron, EventBridge), monitoring automation, or CI/CD pipelines are configured to restart the instance automatically.