The CPUUtilization metric in CloudWatch measures the percent of an EC2 instance’s compute units that are in use. Blue Matador automatically monitors the CPU Utilization of all EC2 instances in your account and alerts when an instance is near 100% utilization.
Since CPU Utilization is measured as a percentage of compute units, it may not match up with the CPU displayed in your EC2 instance’s operating system on instance types that do not get the performance of a full CPU core such as a t2.nano.
When CPU Utilization is at or near 100%, attempt to use SSH to access the EC2 instance and then use tools such as top to get an idea of what is causing high CPU utilization. Check access logs for your application if they are available, and see if any recent code changes could have caused an issue. It is possible that the SSH daemon is unable to accept connections if CPU is truly maxed out. In this case, troubleshoot the instance externally.
If you can find the cause for high CPU on the EC2 instance itself, you can terminate that process or isolate the machine for further debugging by removing it from your Load Balancer. If you are unable to troubleshoot, it may be time to replace the instance or stop then start the instance to get it on a new physical machine.