Monitoring Horizon Performance

[vmmedmed]

Is there a means to monitor Horizon client performance over time? The goal would be to

know about poor VDI experience before or at the same time as the users notice

VDI isn't performing up to its usual speedy self. ?

[dmuligan]

Interesting question, I guess that it could be monitored with vROPs for Horizon but on the Horizon Agent side not on the Horizon Client. With vROPs for Horizon you can measure user experience.

There are default dashboards included in vROPs and also you could create alarms to get notified when some value, for example blast latency, goes beyond a limit.

[JanTrautmann · Accepted solution]

Hi,

for that are 3rd Party tools available. One of them is Riverbed Aternity there you can define your business critical stuff. For example you can track how long a start of an Application was, how the application is preforming. You can define for example the messuring of certain steps, like how long does it take to open this dialog, or when i press this button how long does it take until the next window appears etc etc. And there you get then an overview about why it took so long

Here is an example with excel (only launch is monitored) but you see that for example the server 52.114.6.47 had a high response time which caused excel to start slower.

This here is an example for outlook  :

Beside riverbed are a lot of tools for that like UberAgent etc just search for end user experience monitoring and start testing :=)

With Vrops you get only information about latency, cpu ram which also indicates some information. But when your user calls you and say "Listen today my excel starts terrible slow" and you see cpu.ram hdd etc all okay you have no further information.

With such tools you can extend over the network and servers as well so would see things like that : user starts app -> request send to server -> request got by server after x ms, server had to wait cause was working on a different request etc. Or if you have load balanced services you would see if one of these servers is not okay cause each connect of the users is x ms slower than on all others.

Also such tools have a history, and you can compare changes like installing the 3rd januar update of office365 caused word to take 2 secs longer to start etc.

[vmmedmed]

That's really helpful. We have dozens of remote office and reports of performance vary greatly.