Pod Level Resources Downward API has two test cases. In each
test, a Pod is created with only CPU limits of 1250m specified.
Since CPU limits are set, the CPU requests also default to 1250m.
With a Node capacity of 2000m, two 1250m Pods cannot be scheduled,
which appears to cause the OutOfCpu error.