Omnissa Horizon (on-prem)

The on-prem version of Omnissa Horizon runs in an organization’s own data center or private cloud. This allows enterprises to centralize control of virtual desktops while leveraging their existing infrastructure investments. It is commonly used by organizations that want full control over their VDI environment, data, and compliance, while maintaining flexibility in how virtual desktops are provisioned and managed.

Additional device-level performance data collection is not yet supported for Omnissa Horizon sessions. Therefore, neither the Collector nor the VDI Client Extension installed on the endpoint will send auxiliary information for Horizon sessions.

1. Set up the connector

Configure Nexthink to connect to Omnissa Horizon APIs for session metadata ingestion. This step ensures that contextual data is available to enrich insights within the VDI Experience module.

To set up the connector, perform the procedure described in .

2. Install Collector on VMs

Manual or automated deployment

• Shared VDI: Install the Nexthink Collector on Horizon multi-session hosts. Ensure that session hosts are domain-joined and provisioned with appropriate sizing.

• Personal VDI (persistent): Install the Nexthink Collector as a standard agent during provisioning or onboarding of the persistent desktop.

For both scenarios, perform the procedures described in the following documentations:

1. Configuring Collector for VMs

2. Installing Collector on VMs

3. Deploying Collector

Golden image-based deployment

Pooled VDIs (non-persistent) use a golden image for deployment.

Perform the procedures described in the following documentations:

1. Configuring Collector for VMs

2. Installing Collector on VMs

3. Deploying Collector within a Windows reference image

3. Test the connector and Collector

Connector

Confirm that device and session data is being enriched with information from the Connector integration by verifying fields such as Disk Image, Desktop Pool, Virtualization Host Name, and Virtualization Type. To test enrichment, ensure that the Collector is installed on the virtual machines and is operational.

To do this, run the following query in the Investigations module:

devices
| where virtualization.desktop_broker == horizon_on_prem
| list name, hardware.type, virtualization.disk_image, virtualization.desktop_pool, virtualization.environment_name, virtualization.hypervisor_name,  virtualization.hostname, virtualization.type, virtualization.last_update
| sort virtualization.last_update desc

Collector on VMs

If needed, contact Nexthink support or your Nexthink representative.

Shared VDI

Ensure data such as Device ID, User Name, and Session Hostname is visible in Nexthink.

To do this, run the following query in the Investigations module:

vdi_sessions
| where device.virtualization.desktop_broker == horizon_on_prem and vdi_session.virtualization_type == shared
| list vdi_session.name, user.name, device.name, vdi_session.hostname, desktop_pool, first_seen
| sort first_seen desc

Pooled VDI (non-persistent)

Confirm that instances report unique Device IDs, accurate Pool Names, and fresh Boot Times after deployment.

To do this, run the following query in the Investigations module:

devices
| where virtualization.desktop_broker == horizon_on_prem and virtualization.type == pooled
| list name, virtualization.desktop_pool, boot.last_full_boot_time
| sort boot.last_full_boot_time desc

Personal VDI (persistent)

Verify consistent reporting of Device ID, Assigned User and system uptime across Reboots and Logons.

To do this, run the following query in the Investigations module:

devices
| where virtualization.desktop_broker == horizon_on_prem and virtualization.type == personal
| with session.vdi_events during past 168h
| compute number_of_sessions = vdi_session.count(), time_since_session_last_seen = end_time.last().time_elapsed(), last_user = user.name.last()
| include device_performance.boots during past 168h
| compute number_of_restarts = number_of_boots.sum()
| include session.logins during past 168h
| compute time_since_last_login = time.last().time_elapsed(), number_of_logons = number_of_logins.sum()
| list name, last_user, number_of_sessions, time_since_session_last_seen, number_of_logons, time_since_last_login, number_of_restarts, boot.last_full_boot_time.time_elapsed(), last_seen.time_elapsed()
| sort time_since_session_last_seen desc