Creating a HorizontalPodAutoscaler  in Workloads->HorizontalPodAutoscalers allows you to create an HPA for a deployment that does not have any CPU limit. However, adding one via Workloads->Deployments->[Deployment]->Add HorizontalPodAutoscaler results in an error declaring that a limit is required.

However, this required limit is *not* documented (documentation mentions only needing Requests), and the HPA uses Requests not limits for the scaling logic.
    

Tested on 4.16
    

Reproduced customer's situation in my lab.
    

    1. Create deployment without limits
    2. Click "Add HorizontalPodAutoscaler" from the "3-dots" menu on the right
    

"CPU resource limits must be set if you want to use CPU utilization. The HorizontalPodAutoscaler will not have CPU metrics until resource limits are set."
    

Since we do not use limits anyway, I would expect limits to not be required