Windows 2016 Processor % Graph

[kmitacct]

I have setup a new installation of Cacti to help monitor our Cisco routers and a half dozen Windows 2016 Hypervisors. For the Windows 2016 servers I've tried using the templates for CPU Usage, CPU Utilization, CPUTotal (and individual CPU's), processes, and value 1.3.6.1.2.1.25.3.3.1.2 per some MIB info I read about MS Servers. None of these will give me a % usage for the current usage on the processor.

While looking online for information I keep finding old threads from years ago like this one from 2008: viewtopic.php?f=12&t=29832&hilit=informant+cpu. Does anyone have any information on how to properly graph processor % used on Windows server 2016?

[TheWitness]

Use the Host MIB version of this template. It'll give you all the CPU's. It's in the latest Windows Device Package pretty sure.

[kmitacct]

Thank you for the response. HMIB wasn't installed by default. I had to get it from GitHub. You can download it here: https://github.com/Cacti/plugin_hmib

I'm not too savvy with Cacti so I cleaned up the install directions. I did this on Debian 10. I'm just now getting graphs created and hope it works well.

My directions are below:
1. Untar the package to the Cacti plugins directory, rename the directory to 'hmib'
Plugin Dir Debian 10: /opt/cacti/plugins

2. From Cacti's Plugin Management interface, Install and Enable the pluign.
Console Tab -> Configuration -> Plugins

3. Once you have installed the hmib plugin, you need to install it's Device Template
Console Tab -> Import/Export -> Import Templates -> Imports through browser

4. You must copy the script server script files to Cacti's scripts directory. Those files are located in the plugins templates/scripts directory.
Source: /opt/cacti/plugins/hmib/templates/scripts/
Destination: /opt/cacti/scripts/
cmd ran from source: cp -v ss_hmib_* /opt/cacti/scripts/

5. In addition, you should copy the script server resource files to Cacti's resource/script_server directory. Once there, graphing for Cacti's build in Host Resources Mib objects will come from the hmib plugin.
Source: /opt/cacti/plugins/hmib/templates/resource/script_server
Destination: /opt/cacti/resource/script_server/
cmd ran from source: cp -v *.xml /opt/cacti/resource/script_server/

6. Once everything is in place, you need to goto Cacti's Settings page and locate the 'Host MIB' tab and complete the hmib's setup. From there, you can set collections frequencies and levels of parallelism.
Console Tab -> Configuraiton -> Settings -> Host MIB Tab
Enable "Host MIB Poller Enabled"
I disabled auto discover and purge. Left everything else default.

7. Once you have installed and enabled the Plugin, ensure that you import the Device Package Host_Mib_Summary_Device.xml.gz that is located in the hmib/templates directory.
Must be done from CLI
import package locaction = /opt/cacti/cli/import_package.php
php import_package.php --filename="/opt/cacti/plugins/hmib/templates/Host_Mib_Summary_Device.xml.gz"

[kmitacct]

Using HMIB did not give me the % processor graph I was looking for. In the HMIB tab I can see some % processor data but its not correct. There's no total % processor usage. I'm trying the app processor usage and its givnig me values in processor time. This is part of my issue. I'm looking to graph % CPU like I do Cisco devices but it looks like all the server stuff gives a value that's not 0%-100% and each server has a different baseline so I can't come up with a standard that says value .6 = 25% cpu usage.

I'm starting to think that the best way to grab this data is to use MIB 1.3.6.1.2.1.25.3.3.1.2. It returns a value for each core as a 1 minute average of load. I'm having trouble finding good info on how to build this. I know Perl so I could write a simple script to walk the server with that value and add the total together. I just don't understand much about building custom stuff like this in Cacti.

As a side note I run Zabbix on this server. I've run into issues where Zabbix stopped graphing and after a full migration to a new server with v5.2 the problem persists. I switched to Cacti as the replacement for Zabbix graphing. I ended up testing the Zabbix agent and it did the exact same thing that the Cacti CPU % graphing does. It gives me values that aren't 0%-100% and each server has a different baseline. I'm up for trying anything that will give me a proper value for CPU % used but can't seem to get good data from anything I've tried.

[levansang8689]

So sorry about my post that no relate to this post, however i get the problem with cpu graph for windows server 2019.
please help me to configure to show the cpu for windows server 2019.
i have already done for setting up snmp into windows server.
Thank you!