As per [DOC][1], enabling monitoring dashboards for hosted clusters by setting --monitoring-dashboards --metrics-set=All in the hypershift-operator-install-flags ConfigMap works as expected. After enabling, I could see dashboards created in the Management Cluster for my hosted clusters.
    

However, when disabling monitoring by removing the flags, the dashboards continue to exist (expected) and show data (not expected).

    OCP 4.18
    MCE 2.8.3

    100%

1. Enable monitoring dashboards by creating the ConfigMap:

apiVersion: v1
data:
  installFlagsToAdd: "--monitoring-dashboards --metrics-set=All"
  installFlagsToRemove: ""
kind: ConfigMap
metadata:
  name: hypershift-operator-install-flags
  namespace: local-cluster


2. After a few minutes, Dashboards appear as expected in the management cluster:

Hosted Control Planes / clusters / kubevirt-hcp
Hosted Control Planes / clusters / hosted


3.Disable monitoring dashboards by updating the ConfigMap to remove the flags:

apiVersion: v1
data:
  installFlagsToAdd: ""
  installFlagsToRemove: ""

or explicitly:

apiVersion: v1
data:
  installFlagsToAdd: ""
  installFlagsToRemove: "--monitoring-dashboards --metrics-set=All"


Observe that dashboards for the hosted clusters still receive data.

Dashboards remain in the Management Cluster and continue to display data, even after disabling monitoring through the ConfigMap.
 

When the flags --monitoring-dashboards --metrics-set=All are removed (or added under installFlagsToRemove), the dashboards for hosted clusters should be disabled and stop showing data.
 

Impact:

Users cannot disable monitoring dashboards once they are enabled.

Leads to unexpected monitoring overhead and inconsistency with the documented behavior.

References:
[1] https://docs.redhat.com/en/documentation/openshift_container_platform/4.18/html-single/hosted_control_planes/index#hosted-control-planes-monitoring-dashboard_hcp-observability