Exercise 2: Target Tracking Scaling at 85% and 30% CPU — Possible Solution
====================================================================

A target tracking policy set to keep average CPU utilization at 60%
continuously compares the real, current metric against that target
and adjusts capacity in whichever direction closes the real gap.

At 85% CPU utilization, actual usage is well ABOVE the 60% target --
each existing instance is working harder than the target intends, a
real sign that current capacity is insufficient for the present load.
The ASG will launch additional instances, spreading the same total
workload across more instances until average CPU utilization comes
back down toward the real 60% target.

At 30% CPU utilization, actual usage is well BELOW the 60% target --
existing capacity is doing far less work than the target allows,
meaning some real, unnecessary cost is being spent running more
instances than the current load actually needs. The ASG will
terminate some instances, concentrating the same workload onto fewer
of them until average CPU utilization rises back up toward 60%.

ANSWER: At 85% CPU, the ASG will scale OUT (add instances) to bring
average utilization back down toward the real 60% target. At 30%
CPU, the ASG will scale IN (remove instances) to bring average
utilization back up toward that same target -- in both cases, moving
capacity in whichever direction closes the real gap between the
current metric and the configured target.

WHY THIS WORKS AS AN ANSWER
------------------------------
This correctly explains target tracking as a continuous, two-directional
mechanism (it scales both up and down toward the same target), rather
than describing it only as a one-way "add instances under load"
policy, and ties each real numeric example back to the specific
target the policy was configured with.