Purpose of Exercise 4.3: Resource Limits for a Namespace

crd477

I'm confused about the purpose of this lab exercise. The introduction says:

The previous steps set limits for that particular deployment. You can also set limits on an entire namespace. We will create a new namespace and configure another hogdeployment to run within. When set hog should not be able to use the previous amount of resources.

This seems to contradict step 12 which says:

12. Look at the top output running in other terminals. You should find that both hog deployments are using about the same amount of resources, once the memory is fully allocated. Per-deployment settings override the global namespace settings. You should see something like the following lines one from each node, which indicates use of one processor and about 12 percent of your memory, were you on a system with 8G total.

And this is, in fact, what I observe in my own environment. However, why should they be the same? Isn't the exercise supposed to demonstrate some resource limitation in the low-resource-range case? Shouldn't that one be limited to 500Mi or something?

Maybe a better question is, what is the value of such namespace range limits if the deployment can just override them?