Assume the following deployment specification $-$

$5$ manually created EC$2$ instances already exists

Load balancer and a target group $($TG$)$ with the $5$ instances

The TG is attached with an autoscaling group $($ASG$)$ and scaling policy is based on CPU

The ASG has been created with "minimum" instance of $2$

What is the minimum number of EC$2$ instances that will be active at any point of time in the TG even if CPU thresholds are not breac