Multi Core CPU monitoring WMI and Cacti 1.1.12

[maxcordan]

Hello colleagues!


I had a problem with the template for the CPU monitor on the server for each core separately. The template should display every 24 kernels as a separate line on my server.

Now I see the following:

сфсеш пкфзр.PNG (99.05 KiB) Viewed 1340 times

The data source in the graph is as follows:
PercentProcessorTime0 -> Total
PercentProcessorTime1 -> CPU1
PercentProcessorTime2 -> CPU2
PercentProcessorTime3 -> CPU3
...
...
...
PercentProcessorTime24 -> CPU24

After I turn on debug mode for this device:

cacti log.PNG (230.08 KiB) Viewed 1340 times

(Yes, I create 2 graphs for monitoring single CPU, but I see the same outputs on the graph)

After I set parameter in "Generic Log Level" as "DEBUG" I see the following:

cacti debug.PNG (81 KiB) Viewed 1340 times

(I do not have charts with more than 10 cores (except for this), so I can conclude that this part of the debug applies to this graph)

I have installed Cacti 1.1.12.

In Cacti 1.1.5, a similar problem was not observed. All 24 cores were recovered. Judging by the output debug, I can conclude that Cacti "forgot" to read long lines. I checked that the number of characters with the output of PercentProcessorTime14: * inclusive, is 505 characters.

In Technical Support I have following:

cactits1.PNG (39.97 KiB) Viewed 1340 times

cactits2.PNG (109.43 KiB) Viewed 1340 times

Please help with this. I want look all cores on my graph
Are you have any ideas?

[rony]

Does the server in question have the latest snmpd running?

Can you manually query the OID that Cacti is showing NaN values on the graph?

[maxcordan]

Does the server in question have the latest snmpd running?

Can you manually query the OID that Cacti is showing NaN values on the graph?

Thank for answer!

I get information over WMI. not SNMP.