TECHZEN Zenoss User Community ARCHIVE  

Gaps in graphs when monitoring Windows Server 2018 on https (port 5986)

Subject: Gaps in graphs when monitoring Windows Server 2018 on https (port 5986)
Author: Florin Lupean
Posted: 2018-09-20 06:51

I use Zenoss 6.1.2 to monitor Windows servers, but after I installed the needed zenpacks and made the configuration in Zenoss (Configuration Properties - zWinRMPort=5986, zWinScheme=https) I noticed that I have graphs only for a short period of time (5-10 minutes after device modelling). I searched in ZenPython logs and I found this:
 - WARNING zen.MicrosoftWindows: Windows Perfmon failed to stop Get-Counter on xx.xx.xx.xx: User timeout caused connection failure., Windows Perfmon failed to stop Get-Counter on xx.xx.xx.xx: User timeout caused connection failure.
- WARNING zen.MicrosoftWindows: Windows Perfmon failed to stop Get-Counter on xx.xx.xx.xx: xx.xx.xx.xx Attempted to receive data with no shell and/or command id., Windows Perfmon failed to stop Get-Counter on xx.xx.xx.xx: xx.xx.xx.xx Attempted to receive data with no shell and/or command id.
After I encountered this problem I switched zWinRMPort to 5985 and zWinScheme to http and the gaps dissapeared. Any advices? Thank you!
P.S: I've attached screenshots with the gaps.

------------------------------
Florin Lupean
Junior DevOps Engineer
Evozon Systems
------------------------------

Attachments:

Screenshot_from_2018-09-20_13-35-54.png



Subject: RE: Gaps in graphs when monitoring Windows Server 2018 on https (port 5986)
Author: Luke Lofgren
Posted: 2018-09-21 08:01

Are there any events on the Windows server side? 

Do you have blocking timeouts set in ZenPython?

What does a debug collection run show? 

How many devices / windows devices do you have per ZenPython container?

------------------------------
Luke Lofgren
Infrastructure Architect
Acxiom Corporation (home based associate)
Waterford PA
------------------------------


Subject: RE: Gaps in graphs when monitoring Windows Server 2018 on https (port 5986)
Author: Florin Lupean
Posted: 2018-09-24 09:42

"Are there any events on the Windows server side?"
R: No. Actually this servers are also monitored using a Zenoss 4.2.4, and the graphs are displayed normal. We have the gaps problems only with Zenoss 6.1.2.

"Do you have blocking timeouts set in ZenPython?"
R: In /opt/zenoss/ZenPacks/ZenPacks.zenoss.PythonCollector-1.9.0.egg/ZenPacks/zenoss/PythonCollector I found this: parser.add_option(
'--blockingtimeout',
dest='blockingTimeout',
type='float',
default=30.0,
help="Disable plugins that block for X seconds")
I modified the default value to 0 and restarted the services, but we still have those gaps.

"What does a debug collection run show?"
R: I ran this command : zenmodeler run -v10 --collect=IpInterface -d build.zenoss.loc , but I didn't find any errors. The only error I found was in Zenoss.Core logs and sounds like this : WARNING zen.python: Task xx.xx.xx.xx 300 Windows Perfmon blocked for 4.44 seconds in collect, Task xx.xx.xx.xx 300 Windows Perfmon blocked for 4.44 seconds in collect.

"How many devices / windows devices do you have per ZenPython container?"
R: 4 Windows devices and the awkward thing is that the graphs start and stop at the same time on each server.

------------------------------
Florin Lupean
Junior DevOps Engineer
Evozon Systems
Sebes
------------------------------


< Previous
Timezone Monitoring
  Next
Backing up Analytics
>