Pick the data-driven approach.
Half your EC2 fleet sits at <10% CPU 24/7. The right-sizing play
is:
A) Buy more of the same instance "for headroom".
B) Use Compute Optimizer / CloudWatch metrics to identify
underutilized instances and downsize (smaller instance type or
fewer replicas), revisit weekly.
C) Migrate everything to the largest instance available.
D) Do nothing — overprovisioning is "safer".Sign in to save your code and track progress across devices.
