A few questions...

Hi Jonathan,

That is correct, with 5.6.3 you will have to rerun a discovery for any new servers. The good news is we've just released 5.6.5. In this version we've added a number of command line utilities that let you manually add servers individually without having to rerun the entire discovery. From vWorkspace 8.0, GA next week, and with 5.6.5, servers will be automatically monitored as you add them in the vWorkspace console. The limitation here is for SSL Gateways, VMware vCenters and MS Hyper-V servers, though they can be manually added using the aforementioned command line utility.

In the discovery wizard, if you use "Automatic load balancing", we use a round robin algorithm to target *all* servers to the selected FglAMs. In the case where you rediscover an already balanced environment the preferred method is to use the "Manual selection" option. In this dialog you can search for _newly_ discovered servers and target them to the FglAM of choice.

As I've said, in 5.6.5 you can use the command line utility - documented in the 5.6.5 Admin guide - to add and target a new server *without* having to rediscover:

fglcmd -usr foglight -pwd -srv -port -cmd f4vd:monitorComponent -componentName -componentType -farmName -hostName -remoteClientID -format 0

To get a list of remote "Client ID"s you can use the following command utility: 

fglcmd -usr foglight -pwd -srv -port -cmd agent:clients

A note of warning for if/when you upgrade to vWorkspace 8.0, concerning automatic monitoring. As your use case is not purely load related, that is it also contains a geographical location, you may not get the expected result. More precisely, we do not currently let you choose a FglAM from the vWorkspace 8.0 console. We will automatically choose one depending on the current load *only*. In your specific case, this may not be where you want it to go. If this happens you will have to manually move the monitored server. You can do this via the command line utilities - first remove, then add:

fglcmd -usr foglight -pwd -srv -port   -cmd f4vd:removeComponent -componentName -componentType -farmName -format 0

Now use f4vd:monitorComponent as above.

As for blackouts, unfortunately, I don't think there is an in-built way of blacking out rules for a dynamically changing set of objects. The "Blackout Wizard" uses object groups to statically _filter_ objects. This only narrows the search during the Wizard’s process. To dynamically blackout alarms on objects we'd have to do something customized. One way could be to create a simple rule (time or schedule driven) that checks if all objects of a group are blacked out and if not then black them out. The problem is, it is a non-trivial piece of code to develop - definitely doable though.

You've found a bug. The numberActiveUsers derived metric should be farm specific. I've looked at the expression and it's wrong. I'll raise a ticket. It should read like this:

count(#active from VDIUser where $object.VDIModel.uniqueId = $scope.uniqueId#)

How did you create those tables? Via the drag&drop "Create dashboard..." action? It's difficult to know what's happening without a closer look.

Mark.

I understand what you are saying with the agent assignment and I'll poke around with the command line tool some. Though I can't help but feel that it would be a good idea automatic load balancing worked not just on the who list but on the systems selected.

We are currently vWorkspace v7.5 and working on upgrading to 7.6 so I don't see 8.0 comming very soon here.  I'll install 5.6.5 on our test Foglight server and see how it goes.

As to the report if I remember correctly I used one of the wizards to create the tables here are the details, nothing complex here.

I've created a feature request for load balancing selected servers on selected FglAMs.

There are a number of limitations when creating Dashboards using the Drag&Drop feature. You can acheive what you want, but you'll have to manually modify the row tables.

First you have to "share" the table. Click in the top right of the table and choose "Share..."

In the resulting popup, check at least the "Additional Purpose" "Pagelet", optionally any Relevent and/or Allowed rows, then "OK".

Go to the dashboard configurations via, Configuration>Definitions:

Select the "Row-Oriented Table" that you have just shared:

Click "Edit" to modify the view - NOTE: your Foglight user must have the "Cartridge Developer" role to do this - otherwise this button is greyed.

Go to the configuration tab and make sure "Show Advamced Properties" is checked

Edit the "Default Sort Column ID" property and add the ID of the column on which you would like to sort and click "Set" - in your case that would be c_2 if you want to sort on the current session count.

You can also change the columns sort order. Expand the Columns property then the Column (e.g. c_2)

Change the columns sort order if necessary

Save the view and test.

Thanks for the request and help!

Neat trick, thanks... though I still have the problem that although you can click on the columns and tell it to "sort" at least for the current user sessions count it does not appear to be working correctly (this includes the default sort you had me set), it does in fact change the order of the rows in a consistant maner so something is being sorted but it is not the value being shown. (And as far as I can tell this is not some string value sort vs. numeric value sort as the value shown.)  This also appears to be an issue for the filter as well since I have a filter setting it to only show rows where the average count is greater than 0 and I see plenty of 0.00 values.

Sorry I hate being such a pest on this

I can't seem to reproduce what you're seeing, for both the sort and filter, but I am using 5.6.5, so we're not really comparing apples with apples.

How about if you change /sessionCount/current (MetricValue) to /sessionCount/current/average (Double) in your "Current Sessions" column.

You can use Configuration>Definitions in the same way as before. This time edit the Columns->Column->Value property. You'll have to set the renderer to (none) as the "Average" renderer will only work on metric values.