[Kubernetes Autoscaling] Docs don't explain how to fully disable autoscaling or warn about billing from existing DatadogPodAutoscaler objects

[HassData]

Documentation Page

https://docs.datadoghq.com/containers/autoscaling/

Issue Description

The Kubernetes Autoscaling setup page covers how to enable the feature but provides
no guidance on how to fully disable it. This creates two problems for customers who
want to turn off autoscaling:

1. The page doesn't mention that setting the feature flags to false in the agent
   config is not sufficient to stop billing:

   features.autoscaling.workload.enabled: false
   features.autoscaling.cluster.enabled: false

   These flags prevent new DatadogPodAutoscaler objects from being created, but any
   existing objects in the cluster continue to generate Kubernetes Autoscaling vCPU
   billing until they are explicitly removed.

2. There is no "Disable autoscaling" section documenting the full teardown steps:

   • Set the feature flags to false in the agent config
   • Remove existing DatadogPodAutoscaler objects:
     kubectl delete datadogpodautoscaler --all --all-namespaces
   • To verify, navigate to the Autoscaling page and click "Settings" in the top
     right corner — this shows which clusters still have autoscaling enabled.

Evidence: A support case where a customer disabled the feature flags and continued
seeing unexpected Kubernetes Autoscaling vCPU billing due to remaining CRDs.

Suggested Fix

Add a "Disabling Kubernetes Autoscaling" section to the setup page covering all
three steps above, with a clear note that billing continues as long as
DatadogPodAutoscaler objects exist regardless of the feature flag state.