Introduce API that would allow Dashboard get resources info about target K8s namespace for workspace

[sleshchenko]

Description

It's needed for #13484.
Introduce API that would allow Dashboard get resources info about target K8s namespace for workspace (name, quotas for now).

In the scope of this issue, it's needed to:

• investigate closer Resource API of Kubernetes and OpenShift; It's already done. more see Introduce API that would allow Dashboard get resources info about target K8s namespace for workspace #13504 (comment)
• implement API (or extend /kubernetes/namespace) that would show quotas information about namespace, like RAM, CPUs, PVCs count, etc.;

Additional context

 oc describe quota object-counts                                                                          
Name:                   object-counts
Namespace:              skabashn-che
Resource                Used  Hard
--------                ----  ----
persistentvolumeclaims  1     2
replicationcontrollers  0     20
secrets                 18    20
services                0     15

Name:       compute-resources
Namespace:  skabashn-che
Scopes:     NotTerminating
 * Matches all pods that do not have an active deadline. These pods usually include long running pods whose container command is not expected to terminate.
Resource       Used  Hard
--------       ----  ----
limits.cpu     0     16
limits.memory  0     8Gi

[sleshchenko]

Is suspended in favor of #13538

[sleshchenko]

There is a dedicated issue to provide a list of available namespaces #14376. And this issue is more about extending that API and provide resources information about namespaces. The description is updated with the corresponding changes.

[sleshchenko]

Here is a result of quick research of K8s/OS Resource API I did some time ago. It works pretty in the same way for K8s and OS except that fact that K8s dashboard has its own API proxy that responds with like grouped data (below will be an example and maybe we should do the same for our API).
In the cluster there maybe be different quotas objects (although docs says The administrator creates one ResourceQuota for each namespace See [1]).
Each quota is independent of others like RAM nor pod count limit are not added in the end. If there is one quota with 1 pod, and another quota with 2 pods - then user is able to create only 1 pod(not 3), since otherwise, he will exceed the first quota.
Only admins are able to create quotas - but all users should be able to view them(in namespaces where they has rights) along with status.
K8s API also provides a status for quotas - so, we should not evaluate by our own how much resources are left.

Quotas examples:

---
apiVersion: v1
kind: List
items:
-
  apiVersion: v1
  kind: ResourceQuota
  metadata:
    name: compute-resources-1
  spec:
    hard:
      pods: "1"
      limits.memory: 1Gi
-
  apiVersion: v1
  kind: ResourceQuota
  metadata:
    name: compute-resources-2
  spec:
    hard:
      pods: "2"
      persistentvolumeclaims: "5"
      limits.memory: 2Gi

The error message when user tries to exceed quota:

And K8s namespace proxy API:
http://127.0.0.1:36011/api/v1/namespaces/kube-system/services/http:kubernetes-dashboard:/proxy/api/v1/namespace/test

As we can see, K8s Dashboard receives quotas information along with other kinds of namespace information. And OS Dashboard receives bare namespaces and then bare quotas list.
I've put it here since we may want to have one Che K8s Namespace API that will expose resources info as well, or go in OS way and have Namespace and Resources API separately.

P.S. It's not completed an investigation, like there is limits field in Namespace object of K8s Dashboard Proxy API, but hope it will help a person who will work on this issue.

[che-bot]

Issues go stale after 180 days of inactivity. lifecycle/stale issues rot after an additional 7 days of inactivity and eventually close.

Mark the issue as fresh with /remove-lifecycle stale in a new comment.

If this issue is safe to close now please do so.

Moderators: Add lifecycle/frozen label to avoid stale mode.