Hi,
As per my testing, these 2 commands give a difference utilization of quota limits :
#oc create quota example1 --hard=pods=4 => this quota will be against only "running" state pods inside a namespace
#oc create quota example2 --hard=count/pods=4 => this quota will be counted against any "status" pods inside a namespace. Even status "completed" pods are counted here and I suppose all the pods with any status, except "Failed" ones, will be counted here.
If I have in the current namespace (#oc get pods 1 pod in "running" status and 2 pods marked as "completed" status, then #oc get quota will print :
- example1 quota limit : 1/4
-example2 quota limit : 3/4
In my example, if I understand correctly, example1 is a compute resource quota and example2 is a object count quota of the same resource "pods".
But I've observed not all resources are "countable". So, if I'm replacing resource "pods" with resource "cpu", then the cpu is only limited (appears 20mi/2 is used), but not counted (appears 0/2 used), even if I can create a cpu count object quota.
Can I suppose that for any resource I can create both type of quota : limited quota and count object quota?
