Guidance for monitoring VMs

Introduction

License requirements

VMware/Hyper-V monitoring further details

OS monitoring further details

SNMP monitoring further details

Matrix of metrics provided by each monitoring type

Introduction:

There are several business, financial, and functional factors to consider when deciding how to monitor your virtual machines. In Entuity, there are two recommended methods by which to monitor VMs, each with their own considerations - the managed object in question is a different type in each case, and therefore there are different attributes, associations and metrics available for monitoring. They are independent methods, meaning that you can select either, or run both at the same time.

• VMware/Hyper-V monitoring of VMs, managed as an associated asset to a VM platform.
• OS monitoring of server OSs hosted on VMs (via WinRM or Linux SSH), managed as a server asset.
  • Whilst Entuity usually recommends using WinRM for Windows servers and SSH for Linux servers, you can also monitor Windows and Linux servers on VMs using a suitable SNMP agent. In this case the server would then be managed as managed host, which is classified as a network device.

Please see the below section for a matrix of the metrics provided by each monitoring type.

License requirements:

Each option requires a different license and provides data for different metrics (although there are some overlaps):

VMware/Hyper-V monitoring of VMs:

• Requires a basic device license credit per VM.
• You can pick and choose which VMs you want to manage under license (and you will incur a charge for those licenses). For example, if you have 100 VMs but only want to manage 10 of them, Entuity enables you to select those 10 and leave the remaining 90 unmanaged, and you will be charged for those 10 licenses.
• (The VM platform controlling the VMs requires a full device license. Note, the first onboarded hypervisor on any VM platform does not consume any additional license credit, because it is considered to be covered by the VM platform's full device license credit. The net effect of this licensing scheme is that a vCenter does not cost you anything from a licensing standpoint, but the hypervisors are charged a full device license credit each.)

OS monitoring of VMs:

• Requires a full device license credit per server OS.
• (The VM hosting the server OS does not need to be monitored by Entuity.)

Running both VMware/Hyper-V monitoring and OS monitoring of VMs:

• Requires both a basic license credit for the VM and a full license credit for the OS. There is also the full license credit cost of each hypervisor.

If you choose to run both VMware/Hyper-V monitoring of a VM and OS monitoring of the server OS on that VM, you will be charged for both the requisite respective basic license credit and full license credit (plus the full license credit cost of each hypervisor).

SNMP monitoring of managed hosts:

• Requires a full device license credit per managed host.

VMware/Hyper-V monitoring further details:

Which hypervisor the VM is running on:

VMware/Hyper-V monitoring of VMs tells you which hypervisor each VM is running on, which is not available through OS monitoring.

When VMs move from one hypervisor to another:

With VMware/Hyper-V monitoring, Entuity will open the Virtual Machine Moved incident to tell you that the VM has moved from one hypervisor to another. Such moves can be controlled by the administrator, but sometimes are automated for the purposes of load balancing or to remove machines from the hypervisor in order to shut it down for maintenance. In these cases, these moves are invisible to the OS and so would not necessarily be visible to the user.

Hypervisor history for VMs:

VMware/Hyper-V monitoring provides a full hypervisor history for each VM, detailing the hosting hypervisor and timestamp, and if specified highlighting any changes.

To view the hypervisor history for a VM:

1. Navigate to the Attributes dashboard of the specified VM. Select 'Hosting Hypervisor' from the table and click Show History at the top of the page (or via the Overflow Menu or right-click Context Menu).
   
   
   
   
2. The Hosting Hypervisor on... table will open for the specified VM, with timestamped data of the hosting hypervisor in 5 minute intervals.
   
   
   
   
3. Click Show Changes Only at the top of page (or via the Overflow Menu) to view details of when the hosting hypervisor was changed.
   
   

Clusters, datastores and hypervisor metrics:

From Entuity v21.0 upwards, VMware monitoring provides details of VMware clusters, datastores, and host bus adapters (HBAs), details of which are invisible to OS monitoring of the VM.

Please also see this article for information on the system dashboards covering these metrics.

VMware monitoring also provides data across all hypervisors, e.g. hypervisor total memory, which again are invisible to the OS. Please see the Associations section of the matrix of metrics below for further details on monitoring VMware clusters, datastores and hypervisors

OS monitoring further details:

If you want to monitor OS services on a VM, either Windows or Linux, you must do so through OS monitoring. The OS metrics (architecture, description, distribution, distribution version, domain name, hostname, manufacturer, name, release, type, version, and version code name) are invisible to VMware monitoring, and therefore available only through OS monitoring.

SNMP monitoring further details:

Whilst Entuity does not recommend this as your first choice (and instead recommends monitoring Windows servers via WinRM and Linux via SSH), you can also monitor Windows and Linux servers on VMs via a suitable SNMP agent. In this case the server would then be managed as managed host, and therefore would be considered a network device rather than a server. Note, the Windows SNMP agent is now technically deprecated by Microsoft.

SNMP monitoring of Windows and Linux servers provides only the OS name, but no other OS attributes.

Although storage volumes are monitored via SNMP monitoring, they do not provide functionality to generated threshold-based utilization events.

Matrix of metrics provided by each monitoring type:

Please find below the metrics available for each type of monitoring, including the Associated attributes for each type: