Description of problem:

As shown in https://issues.redhat.com/browse/RHODS-1768, there are some cases in which OpenShift UI/UX quirks could confuse RHODS customers. 

The issue there was that OpenShift UI was reporting enough available resources to spawn a notebook server, however the `oc describe node` command (showing requests/limits) reported over-committed nodes; this results in the user being unable to spawn a server with an error message of 'not enough CPU', although the UI is reporting something else.

https://issues.redhat.com/browse/RHODS-335 should alleviate the issue and prevent the majority of users from ever seeing it, however there might still be a case in which the user does not have enough available resources to spawn any server sizes with the UI reporting otherwise.

I think this behaviour needs to be documented in the triage guide if not done already, and possibly be explained to users as well (either in docs or spawner). 

Prerequisites (if any, like setup, operators/versions):

More of an issue in RHODS versions before 1.1.1-21, might still be a problem in future versions

Steps to Reproduce

1. Install RHODS on a cluster with enough resources to spawn notebook servers of size S/M/L/XL
2. Over-commit worker nodes
3. Try spawning the notebook server for which you theoretically have enough resources
4. Spawning should fail because of "not enough {resource}" where resource is the over-committed resource in the nodes.

Actual results:

Expected results:

Reproducibility (Always/Intermittent/Only Once):

Intermittent, I don't think anyone will over-commit nodes on purpose.

Build Details:

Additional info: