Win32 - Monitor running processes

[BSOD2600]

Is the Re-Index Method set to Index Count Changed or Verify All Fields ?

Where to do that?

In the device, under associated data queries.

[Discogalaxy]

Oh, i must be bilnd
I searched and searched....

Thx for info

nice script by the way

[Dr.J]

Is there an easy way to disregard CPU values of say, less than 1% using a CDEF or similar? How about rounding results to the nearest whole number?

Can anyone point me in the right direction with how I might do this?

Thanks in advance

[BSOD2600]

Is there an easy way to disregard CPU values of say, less than 1% using a CDEF or similar? How about rounding results to the nearest whole number?

Modify the data template and change the min value to 1 (or whatever you want). You'll have to either run rrdtool tune on the associated rrd files or delete them so they're remade.

As for rounding, change how they're displayed in the graph template.

[esimon]

it don't work
Can you help me please
I import all graph
i copy win32_procs.ml
i create data query but where you change re-index
thank you

[BSOD2600]

1) did you add the xml script to the correct folder?
2) did you add the associated template to a windows host?
3) does it return any rows when you click on verbose query?
4) did you go to the 'create graphs' screen and select the processes you wanted graphed?
5) did you wait 10 minutes before looking at the newly created graphs?

[maw]

Hi BSOD2600,

This mod looks really interesting... few questions before I implement.

1)
If my IIS servers have multiple Application Pools how will this mod handle more then one IIS worker process? After reading all 10 pages on this topic it seems the mod grabs the first process it finds and graphs until a reindex is completed. Any ideas to monitor many processes that have the same name? (I read that monitoring on PID will _work_ until the PID changes...you explained the limitations of this approach earlier)


2)

"An inefficient way to fix this is to change the reindex method to verify all fields; that way cacti will reindex it every polling cycle. "

How will performance be affected if I am polling ~150servers every 5mins?

Thanks a lot!

[BSOD2600]

..it seems the mod grabs the first process it finds and graphs until a reindex is completed. Any ideas to monitor many processes that have the same name? (I read that monitoring on PID will _work_ until the PID changes...you explained the limitations of this approach earlier)

Sounds like you've got a good idea how the script works. Correct, since the script uses an snmp xml script to query the hrSWRunIndex and hrSWRunName of computer, we have two options for keeping track of a process. If we go by name, then the script picks up the first instance of hrSWRunName. This works great for unique instances of processes. As you've already outlined (along with other users) the problem arises if one wants to monitor multiple instances of the same hrSWRunName. Since the hrSWRunIndex is actually the processes PID, we have the option to explicitly monitor a process by that instead. Of course then we have the problem of if the process's PID changes (as they do on termination/execution), then the script no longer queries the correct/valid PID. This requires one to manually fix it in Cacti.

I have yet to find a way to fix this while still utilizing SNMP xml queries. If you do, let me know.

How will performance be affected if I am polling ~150servers every 5mins?

Not entirely sure, but it will force cacti to snmpwalk/snmpnextget a large subset of OIDs (snmpwalk .1.3.6.1.2.1.25.4.2.1 to see yourself). You could enable recaching on a host or two, look in the cacti log how long recaching takes and then extrapolate the time cost for ~150 servers.

My slow ass cacti server takes ~2 seconds to recache a single windows host for it's processes, utilizing the cmd.php poller.

[mktoon]

Hello, anyone know how to intergrate this with Thold to monitor a process if it stops and send an email to the admin?

Thanks a lot

[Mika2006]

*
all template i try anyone work
My config :

| Cacti 0.8.6j |
| MySQL ? | PHP ? | httpd | RRDtool 1.2.15 |
| RedHAT 4es | P4 3000Mhz | 512MB |

1 / i import on the cacti web :
cacti_data_template_snmp_running_process_info.xml
cacti_graph_template_snmp_running_proc_info.xml


2/ i copy files here:
/var/www/html/cacti/resource/snmp_queries
-rw-r--r-- 1 1000 users 1847 mai 13 00:27 win32_procs.xml

3/ and in my cacti interface :
in debug mode :

Data Query Debug Information
+ Running data query [12].
+ Found type = '3' [snmp query].
+ Found data query XML file at '/var/www/html/cacti/resource/snmp_queries/win32_procs.xml'
+ Error parsing XML file into an array.
+ Found data query XML file at '/var/www/html/cacti/resource/snmp_queries/win32_procs.xml'
+ Found data query XML file at '/var/www/html/cacti/resource/snmp_queries/win32_procs.xml'
+ Found data query XML file at '/var/www/html/cacti/resource/snmp_queries/win32_procs.xml'

what do you think please?
please help

[BSOD2600]

This script only works against Windows.

[Mika2006]

???
he exist a difference betwen cacti on linux and cacti on windows?

[Mika2006]

he is so good template
thanks too
in a service graph you have :
-CPU usage
-memmory usage

and i need one more information:

i need how many process X running
someone can help me to make it in this template??

[BSOD2600]

i need how many process X running

Cacti already comes with a template for Windows systems for this.

[Mika2006]

yes i knwo but is all process
and i need by process
like this

Code: Select all

 2 process matching EXCEL.EXE (> 0) 

?