We have created two clusters in XCP-ng:

A Management Cluster consisting of 2 physical hosts.
A Compute Cluster consisting of 5 physical hosts.

We deployed the CloudStack management components on VMs running in the Management Cluster.

During HA testing, we observed that when a VM is shut down from the XCP-ng side, CloudStack continues to display the VM as running for approximately 10 minutes before the status is updated.

We have performed extensive validation on the CloudStack side and have not identified any configuration or operational issues. Based on our findings, the delay appears to be related to the synchronization or communication between XCP-ng and CloudStack, resulting in slow propagation of VM state changes from XCP-ng to CloudStack.

We would appreciate your assistance in investigating the cause of this delay and identifying any configuration or performance-related factors that could be affecting the status update mechanism.