Autoscaler's "CPU Limit Target" in UI does not match what the CLI returns

[alikhajeh1]

To reproduce this issue:

1. From the Deployment UI > Actions > Add autoscaler, then set "CPU Limit Target" to 72. Click on save and confirm that the UI shows 72%:

2. Using the CLI, run oc get hpa and notice it doesn't match the UI:

NAME               REFERENCE                           TARGET    CURRENT   MINPODS   MAXPODS   AGE
php-laravel-todo   DeploymentConfig/php-laravel-todo   600%      2%        1         3         14m

oc describe hpa/php-laravel-todo also shows the issue:

Name:				php-laravel-todo
Namespace:			examples
Labels:				app=php-laravel-todo
Annotations:			<none>
CreationTimestamp:		Thu, 25 May 2017 18:00:53 +0430
Reference:			DeploymentConfig/php-laravel-todo
Target CPU utilization:		600% <<<<<<<< not 72%
Current CPU utilization:	4%
Min replicas:			1
Max replicas:			3
No events.

The min and max replica counts match.

$ oc version
oc v1.5.0+031cbe4
kubernetes v1.5.2+43a9be4
features: Basic-Auth

Server XXX
openshift v1.5.0
kubernetes v1.5.2+43a9be4

Tested with Chrome and FireFox.

[spadgett]

@alikhajeh1 When overcommit is configured for the cluster, the web console shows the target as a percentage of CPU limit, whereas the CLI is showing it as a percentage of CPU request. See also

openshift/origin#13907 (comment)

Unfortunately this is confusing if you use both the CLI and web console. @jwforres I think we should look at the user experience here.

[jwforres]

Closing this one since it is a duplicate of openshift/origin#13907