Win32 - Monitor running processes

argon0 · Post by **argon0** » Mon Mar 04, 2013 6:21 am

Phalek, you got the right thing.

If you manually choose Index Type hrSWRunName, and specify the Index Value with a running process name (e.g. sqlservr.exe) the query (and thus the graphs) work... So this MUST be a valid way to do this... but HOW do you get the data template to use this/these values?

I'm going to do some experimentation with the data template - I have an idea...

argon0 · Post by **argon0** » Mon Mar 04, 2013 6:44 am

Well.... that seems to have worked... Instead of going by the manual, which says not to touch these three fields if you change the fields to look like below... it works (or looks like it does - I haven't restarted the proc yet, so can't comment on if it works after a restart).

: View of Data Procs Template as changed by me...; Proc_Data_Template_pic.PNG (13.67 KiB) Viewed 5023 times

You get Data Sources which look like this:

: Data Source seems to be missing the index type and index value fields - but these are specified in the template...; Data_Source_Example_Process.PNG (24.29 KiB) Viewed 5005 times

And creates a graph like this (zoomed in a bit as its only been running a few minutes):

: Graph created after I've changed the data Template; Graph_Created_with_new_Data_Template.PNG (23.13 KiB) Viewed 5023 times

Finally a copy of my data template ...

{editted thanks to Phalek pointing out server names!}

argon0 · Post by **argon0** » Mon Mar 04, 2013 6:44 am

Interestingly this seems to have kick started my other non-working graphs as well! (Though I suspect coincidence, as I did run queries a few minutes before changing the template)... Those that rely on PID are still not working.

BSOD2600 · Post by **BSOD2600** » Mon Mar 04, 2013 1:14 pm

argon0 wrote:Well.... that seems to have worked... Instead of going by the manual, which says not to touch these three fields if you change the fields to look like below... it works (or looks like it does - I haven't restarted the proc yet, so can't comment on if it works after a restart).

Do NOT do this. You're only asking for problems.

For all Data Query templates (which includes [SNMP - Running Processes]), the correct way to create graphs is:
Devices -> select a device -> click on [Create Graphs for this Host] -> select the item(s) from each Data Query -> click Create at the bottom of the page.

argon0 · Post by **argon0** » Mon Mar 04, 2013 2:21 pm

BSOD2600 wrote:
argon0 wrote:Well.... that seems to have worked... Instead of going by the manual, which says not to touch these three fields if you change the fields to look like below... it works (or looks like it does - I haven't restarted the proc yet, so can't comment on if it works after a restart).
Do NOT do this. You're only asking for problems.

For all Data Query templates (which includes [SNMP - Running Processes]), the correct way to create graphs is:
Devices -> select a device -> click on [Create Graphs for this Host] -> select the item(s) from each Data Query -> click Create at the bottom of the page.

Yes, done as in that is exactly the way I create these graphs - what I do above doesn't change that. I am using variables i.e. |query_hrSWRunName| for the index value, the index type needs to be hard coded, or it will default to the first in the list (i.e hrSWRunIndex - it seems to be alphabetical). I'm just changing stuff back at the data template - not changing the process of creating graphs - well I am - but making it so you don't have to go back and edit them after creation to get the data source working properly. Now, whether I have stored up problems for myself later on or not is another question, but I don't see how...

If those fields aren't designed to be used, why are they there?

BSOD2600 · Post by **BSOD2600** » Mon Mar 04, 2013 6:17 pm

argon0 wrote:I am using variables i.e. |query_hrSWRunName| for the index value, the index type needs to be hard coded, or it will default to the first in the list (i.e hrSWRunIndex - it seems to be alphabetical).

Something's not working properly then as

Code: Select all

   <index_order>hrSWRunName:hrSWRunIndex</index_order>

in win32_procs.xml should've made hrSWRunName the default. Just tested some new items on my installation and also ran into the same problem. I'll file a bug. It should have the process name it it:

: cacti - its-sp-rx1 - procexp64.png (22.36 KiB) Viewed 4994 times

argon0 wrote:If those fields aren't designed to be used, why are they there?

They're suppose to be used for data source and graph titles.

argon0 · Post by **argon0** » Mon Mar 04, 2013 6:29 pm

Really?

OK.

Is this due to 0.8.8a Then?

BSOD2600 · Post by **BSOD2600** » Mon Mar 04, 2013 6:53 pm

argon0 wrote:Is this due to 0.8.8a Then?

I haven't created new graphs with this template for quite a while so it very well could be.

argon0 · Post by **argon0** » Tue Mar 05, 2013 8:13 am

The behaviour you describe above is what I expected to happen, so was surprised when graphs broke post a service restart.... Thus the digging.

argon0 · Post by **argon0** » Tue Mar 12, 2013 5:06 am

As it currently stands I can't use these graphs.

Essentially they are fine when I set them up - either with the original template, or with my edit of it, which "hardcodes" |query_hrSWRunName| into the graph. On a restart of a service everything goes screwy though.

With my edit I end up with a load of (broken) graphs with the name HOSTNAME - Process - |query_hrSWRunName|, and (because I used Autom8 to create graphs) a loat of new HOSTNAME - Process - PROCESSNAME graphs - which work until the next restart, with the original I ended up with a load of broken HOSTNAME - Process - .... graphs.

So.... Any idea of a fix to the "bug"? So I can go back to the original graphs?

phalek · Post by **phalek** » Tue Mar 12, 2013 5:36 am

Maybe something like this template would be sufficient for you ( to get availability ) in the first place ?

I'm using it for the service availability calculations in my reporting plugin.

http://forums.cacti.net/viewtopic.php?f ... &start=135

argon0 · Post by **argon0** » Tue Mar 12, 2013 9:43 am

I'm monitoring Services fine, its when I try to do Processes that the problem happens...

BSOD2600 · Post by **BSOD2600** » Tue Mar 12, 2013 5:33 pm

argon0 wrote:As it currently stands I can't use these graphs.

1) dont use automate for this template, as it doesnt work properly.
2) the reindex method should be index count changed for this data query.
3) the data template for each device's instance needs manual updating until the bug gets fixed. Index Value should be the process executable (i.e. spoolsv.exe) and Index Type should be hrSWRunName.

argon0 wrote:So.... Any idea of a fix to the "bug"? So I can go back to the original graphs?

bug the developers to look into http://bugs.cacti.net/view.php?id=2350 I suppose.

JJX · Post by **JJX** » Tue Jul 26, 2016 9:15 am

I have two processes with the same name.
I match the first process with the name; the 2nd one with the PID.

This is the graph of the 2nd process, matched with the PID:

: A.png (69.18 KiB) Viewed 3252 times

The top graph is the one appearing to the Graph tree; it doesn't depict the memory.
However, the realtime graphs shows the memory.

Any idea why this happens?

JJX · Post by **JJX** » Mon Aug 01, 2016 2:02 am

After deleting the rrd file and running again the poller the problem was resolved

Cacti

Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Re: Win32 - Monitor running processes

Who is online