javier.vazquez

List Views observations, questions, and complaints

Discussion created by javier.vazquez on Mar 29, 2011
Latest reply on Mar 31, 2011 by javier.vazquez

Guys,

 

I am working at a prospect that I decided to import the UMP 2.0 On Prem.LAR  file.  This allowed me to leverage all of the list views such as

 

>Summary Views

>>Datacenter

>>>Server Health

 

This particular view has a column called 'CPU Usage'.  The QoS is 'QOS_CPU_USAGE', the Target is '$HOST', and the Display has the settings for how to display.

The issue with this is that when I open this list for viewing only two of the servers actually have bars for memory and cpu and the rest are just blank. 

 

I then went in and edited that list and saved under a different name.  The edit I made is to change on the 'Target' column from '$HOST' to '*'.

This seemed to populate my list almost entirely with gauges for CPU and Memory. The problem is that I'm seeing many CPU's that show a full bar which is red.  In other words the list is claiming that my CPU is at 100%.  For at least 40% of the servers that I look at in that list(BTW: I'm dealing with all physical servers running CDM probe) this is flat out wrong.  This made me think that the setting I changed may be taking all metrics like "idle", "wait", "system", "user" and "total" and took an average or even a sum of all of them.

 

>>The question is:  what is the difference between using '$Host' and '*' for the Target field on a list?

 

Next, I have an environment in which possibly 100 hosts are being pinged from multiple HUBs.  100 Hosts are being monitored for CDM.  The only other probes collecting QoS are SQL, Exchange and a couple randoms not doing much. I have a 16 CPU, 32GB RAM beast of a SQL Server.  When I load lists they take a very long time to load certain list views.  We aren't talking about thousands of hosts, we are monitoring < 100!!

 

>>The questions is: What could the delay be?  With such a database shouldn't my result be close to instantaneous?

 

I was told that we are running several hundred or sometimes thousands of queries per list.  If you leave that page open supposedly the queries are re-run and re-run over and over again.  Not to mention if you make some kind of a change to a list views preferences this will automatically refresh the queries to all of the other list views on that page.  This is an enormous amount of strain on the SQL Server.

 

These lists came out and seemed like such a scalable and logical way to view data.  Now I am regretting having introduced them at all.  I realize we have a Top N option coming out soon.  It would be great to hear more about this and also including a possible release date.

 

Anyone else?  Thoughts?

Outcomes