Version of <application usage summary instance>

(imported topic written by Vosh)

I’m using client property application usage to track the run count of java.exe, but I’m only interested in tracking version 1.6.x (6.0). However clients could have both java versions 6 and 7 installed. I’ve tried using the BES client api tester to figure out how to use the "
version of "
to pull the java version so I can count only the one we need but I’ve only managed to get “version of” not defined.

Ideally I’d like to see (total run count of it, first start time of it, last time seen of it, total duration of it) of application usage summaries “java.exe” include “version of it” so it looks like;

(version of it, total run count of it, first start time of it, last time seen of it, total duration of it) of application usage summaries “java.exe”

But that doesn’t seem to work either, that and I think it’s pulling the run count of it from

"(
item

0

of

it

(
item

1

of

it

0
))
of
application usages
“java.exe”"

Which is useless to me unless there is a way to make it count only when version of running application contains “6.0”.

Does anyone have thoughts on this?

Thanks!

(imported comment written by NoahSalzman)

I had to look up how that inspector works to figure this out. First I did a search for the inspector:

Q: properties whose (it as string contains “application usage”)

A: application usage summary : application usage summary

A: name of : string

A: total run count of : integer

A: first start time of : time

A: last start time of : time

A: last time seen of : time

A: total duration of : time interval

A: running of : boolean

A: application usage summaries: application usage summary

A: instances of : application usage summary instance

A: name of : string

A: total run count of : integer

A: first start time of : time

A: last start time of : time

A: last time seen of : time

A: total duration of : time interval

A: size of : integer

A: string version of : string

A: version of : version

A: application usages : timed( time range, integer )

From that I saw that version had to be used with the “instance” of a usage summary… (there can be multiple summaries per usage report):

A: version of : version

A: instances of : application usage summary instance

Which led me to:

q: (versions whose (it = “6” as version) of instances of it, total run count of it, first start time of it, last time seen of it, total duration of it) of application usage summaries “java.exe”

A: 6.0.300.12, 32, ( Thu, 09 Feb 2012 15:42:29 -0700 ), ( Mon, 18 Mar 2013 11:02:18 -0700 ), 19:54:46.861

A: 6.0.0.0, 32, ( Thu, 09 Feb 2012 15:42:29 -0700 ), ( Mon, 18 Mar 2013 11:02:18 -0700 ), 19:54:46.861

Note that dealing with the version inspector requires the “as version” and quotes around the version number.

(imported comment written by Vosh)

Perfect, the “as version” when using the version inspector was throwing me off. This should work nicely, thanks!

(imported comment written by Vosh)

Actually I do have an additional question. When using this with the application usage summary wizard and editing the analysis created from it, there are three properties created. The last of which I plan to substitute your relevance in, “Application usage summary: java.exe”.

However the first entry Application usage uses the following relevance,

(item 0 of it * (item 1 of it > 0)) of application usages “java.exe”

So my question is when using the relevance you iterated , is it pulling the “count” from the above relevance and thus potentially counting all versions of java.exe processes? Or perhaps I’m missing how these two properties may, or may not, be tied together?

And how often should I set this report to evaluate?

(imported comment written by NoahSalzman)

Hi Vosh,

Be careful, the analysis you are editing is intended only to feed data to the Application Usage Dashboard. So, if you change the analysis or the Computer Properties without also changing their names you will be affecting what the Dashboard displays.

The analysis that you are questioning is used to feed data to the Dashboard. “item 0” of “application usage” will show the duration that the client was measuring/looking for application usage. “item 1” is a count of the times that exe was seen during the measured period.

q: (item 0 of it, item 1 of it) of application usages “firefox.exe”

A: ( Mon, 14 Oct 2013 10:00:26 -0700 to Mon, 14 Oct 2013 10:01:33 -0700 ), 0

A: ( Mon, 14 Oct 2013 10:00:26 -0700 to Mon, 14 Oct 2013 10:01:33 -0700 ), 1

A: ( Mon, 14 Oct 2013 10:01:33 -0700 to Mon, 14 Oct 2013 10:02:07 -0700 ), 0

The analysis is setup to find out if there were any instances found – “item 1 of it > 0” – which returns true or false. That true/false is then multiplied against the “time range” to put true or false next to that duration so that it is more easily consumed by the Dashboard.

q:
(
item

0

of

it

(
item

1

of

it

0
))
of
application usages
“firefox.exe”

A:
( Mon, 14 Oct 2013 10:00:26 -0700 to Mon, 14 Oct 2013 10:01:33 -0700 ), False

A:
( Mon, 14 Oct 2013 10:00:26 -0700 to Mon, 14 Oct 2013 10:01:33 -0700 ), True

A:
( Mon, 14 Oct 2013 10:01:33 -0700 to Mon, 14 Oct 2013 10:02:10 -0700 ), False

I don’t know how the Dashboard consumes all of the analysis data. So I’m not quite sure how to answer your question. I guess what I would ask you is what information are you trying to gather? If you don’t need all the analyses that the Dashboard is setting up then just ignore them.

(imported comment written by Vosh)

That makes sense, though I’m still unclear then on how the “total run count of it” is being determined from the relevance. I want to make sure that what’s being counted is actually java versions 1.6.x. But the “total run count of it” of this relevance suggests that “some” kind of voodoo, or Jessica Lange or something is actually keeping track of that count, somewhere…;.


(versions whose (it = “6” as version) of instances of it, total run count of it, first start time of it, last time seen of it, total duration of it) of application usage summaries “java.exe””

To elaborate, prior to you making the relevance use of the version inspector more clear to me I was counting the number of lines that contained the entry “JRE6” from the measure.log file of the bes client.

"number of
lines
whose
(
it

contains

“jre6”

as
lowercase)
of
files
whose
((name
of

it

=

“measure.log”
)
AND

exists
lines
of

it
)
of
folder
"__BESData__Global"

of
parent folder
of
client"

So out of curiosity I added your relevance to the same analysis as an additional property to compare the two. The results were quite different, which isn’t all that surprising, but at a loss for where it’s getting that run count.

Which brings me back to me querying about (
item

0

of

it,

item

1

of

it
)
of
application usages
"java.exe"
or firefox.exe in your example. Both of these using either application usage, or application usage summary in the relevance, and am I correct (or not) that application usage summary correlates to application usage, and thus the actual count could or total count could be inaccurately counting all versions of said process?

Or am I over-thinking this?

(imported comment written by UrosZ)

I have a question. How to enable application usage summary inspector.

q: (versions whose (it = “14” as version) of instances of it, total run count of it, first start time of it, last time seen of it, total duration of it) of application usage summaries “winword.exe”

E: application usage summary inspector is disabled