Looking to gather more info on Virtual Desktops (other than what it does by default)
If we navigate Foglight for Virtual Desktops -> View the health of my VDI environment -> Terminal Servers.
Open terminal server Explorer.
It has metrics like CPU, memory utilization, Network, Disk along with other detailed info.
But Foglight for Virtual Desktops -> View the health of my VDI environment -> Virtual Desktops
Open in VDI explorer, it shows very limited info, but not as detailed like Terminal Server Explorer.
Looking for more detailed performance info on VDI Explorer as well.
For example: If McAfee app is taking too much memory/cpu/disk on this VDI, it will help to know which App is causing this problem & can quickly fix it right away.
Is this something possible in Foglight for VDI?
If not possible, what could be reason, why this dashboard is not loaded with detailed info like the Terminal server does?
Information about an individual VDI can be found if using 5.6.3 onwards. This release includes vFoglight which can monitor VMWare and HyperV environments - specifically it can provide CPU load, Network I/O, Disk I/O and Memory utilisation information - Disk I/O and Memory are available only to VMWare. To set this type of monitoring up, should run through a discovery - this process should pick up the VMWare/HyperV environments that are known to vWorkspace.
Once the discovery process has been completed, the UI used to show the vWorkspace environment will have two additional options - VMWare and Hyper-V. Using these two tabs the VDI in question can be found and will display relevant information about these installations.
With respect to applications and processes, it's not that we can't collect this data, it's that there are some strong reasons why trying to would lead to some serious consequences. The main factors influencing this are the different characteristics of TS vs VDI.
Characteristics of TS in installations are:
1) Fewer in number (ie: typically, at most, several hundred)
2) Relatively static (ie: aren't generally dynamically provisioned, decommissioned etc)
Characteristics of VDI in installations are:
1) Large numbers (ie: often several thousand)
2) Dynamic - VDI are provisioned, reprovisioned, decommissioned very quickly (often daily)
Why do these characteristics matter?
Well, to monitor the applications/processes of a host we need to use an agent from the Infrastructure cartridge (typically a WindowsAgent type). As we've seen in the past, these agents are resource intensive (to create, configure, maintain and run) and each FglAM can only support approx 250 agents in total. For TS we can create these agents an monitor the applications/processes etc as there aren't very many of them (we only need a handful of FglAM) and they don't keep changing - generally, once a TS is in place, it's not going anywhere soon! This means that once we've carefully deployed the agents across our FglAMs there's little more we need to do.
Contrast that with VDI - we'd need thousands of agents to monitor each VDI, spread across hundreds of FglAMs, and due to the dynamic nature of their provisioning we'd be creating and destroying agents very frequently which would place a huge burden on the FMS. We'd also need some complex algorithms to dynamically handle the creation and deletion and proper load balancing across FglAMs to monitor the ever shifting numbers of VDI.
Aside from the practicalities of how to monitor the VDI, there's another factor to consider, the amount of data that would be generated. If, there were thousands VDI all having data of applications/processes collected it would be a significant amount of data for the FMS to handle which might be detrimental to performance.