Governance_Dashboard - what about the performance data?

[datanibbler]

Hi,

I am building a new Monitoring_Dashboard based on the data in the GD (Version 1.1). I am just covering Performance data and I am not quite sure how the figures interact:

- Allocated RAM is what the Server has made available to QlikView;

- Committed RAM is what QlikView is actually using

- Free RAM I just don't know ...

=> The issue is, I cannot pick two of those three figures and add them up and that is a static total which doesn't move - I have tried: When I add up "Committed" and "Free", that is m.o.l. always the same nr. of GB, the nr. of MB varies.

<='> But the amount of RAM the Server has allocated to QlikView is not "free" anymore, is it? So I should add up "Allocated" and "Free" maybe - that also Returns a total which varies slightly - but if I have a look at the Windows_monitor or whatever it's called (by pressing >> Ctrl+Shift+ESC <<) I see that the total of physical Memory is less than that sum

Can anyone help me here?

Thanks a lot!

Best regards,

DataNibbler

[datanibbler]

The ultimate Goal of that Dashboard, among others, is to identify the capacity we have for any additional Jobs to run on that Server.

To that end I want to Display the amount of free Memory at given times of the day - but to achieve that, I Need some fixed upper Limit that I can use to calculate that remaining capacity.

It seems that the amount of RAM differs rather a lot over time ...

[Tyler_Waterfall]

I have seen one astute customer integrate perf logs into this type of dashboard. That will give you a better picture of the actual resources consumed/available on a server.

Keep in mind that the server logs report the CPU and RAM usage by QV Server process only. This does not include QVB.exe running during reloads which can be significant resource consumers.

[datanibbler]

Hello

what is QVB? I haven't heard that term before.

Performance seems to be really complicated - I have a table "Performance" with those three measures, but they don't seem to really go together.

I guess that "Allocated" plus "Free" should match the total amount of Memory that we have in our VM, no? Well, it doesn't - according to the PerfMon (Strg+Sh+ESC), we have a total of about 132GB -.  but at times the sum of "Allocated" and "Free" is way above that. I guess at those times the Server starts using the harddisk.

The final Goal of that Monitoring is to identify at which times we still have capacities so that we could take on some more Jobs, I don't yet see how that can be done. For the Moment I'll Focus on displaying some more data on the status_quo like the nr. of concurrent users, RAM_usage and so on - I always look at the actual GD to get some ideas.

One more Thing you could maybe help me with: I wonder, does it really make sense, given the granularity of my calendar (which is the central table of my datamodel),  to Show data from the "Performance" table with user_data (from the "SessionTaskAuditMaster") in one Chart? I have my calendar down to hour-intervals - but saying that I had 60 (different) users within an interval of one hour doesn't tell me too much ...

My central Dilemma is this: Increasing the granularity to minutes would of course increase the size of my calendar by 60 which does not seem like a good idea - but anything less will not really be of much value when it Comes to Performance_data ... What would you suggest here?

Thanks  a lot!

Best regards,

DataNibbler