I have an AWS ECS Service running in the Fargate mode. My setup has two tasks running an httpd image (Apache2) listening on port 80. I have an application load balancer that redirects port 80 to a target group. That target group is configured with two IPs (each task exposes on private IP, hence two IPs in the target group.
I have a question around auto scaling on ECS Services: how does the auto scaling will work in terms of assigning IPs to the target group? That is an essential part of of the scaling-out mechanism since if the new task's private IP is not assigned to the target group then that new container/task won't get any traffic, which counters the entire purpose of auto scaling.