With v10 Orion is now identifying our Palo Altos as Palo Altos. However, I still am not seeing CPU and Memory gauges on the web page. Any ideas why not?
It must be using non-standard mibs to pull that info and we would need to customize our polling engine to go after the mibs it uses for Palo Altos. So you will need to use UnDP's to gather this info.
Did you ever create a UDP to collect this info? If so can you export and upload it to the UDP repository on thwack. I want to collect the same data and figured why re-create if you already did it.
We created a UnDP for this in a customer environment. I cannot upload the UnDP for you, but I can tell you that for CPU load we used OID: 1.3.6.1.2.1.25.3.3.1.2.1 in a tabular format. We have as yet been unable to find one that pulls memory from those devices. If anyone has, we would appreciate it.
Hope this helps.
Thanks-
Confirmed working on a PA-2000.
MIB Value RaW
Format None
Type GET
I second the request for memory stats!
Hey, just an update since this seems to be one of the few threads on this issue.
I still can't get NPM to recognize Palo Alto device CPU the way that I am used to (List Resources -> Check Boxes -> Done). They seem to use some alternative set of OIDs and NPM can't figure it out.
I had to create UnDPs for thee like Chrystal suggested, however there was a change in 4.0 of the PA code and some older info is floating around.
See here:
In PAN-OS 3.1, there are three hrProcessorLoad OIDs:
• .1.3.6.1.2.1.25.3.3.1.2.1 for MP-System
• .1.3.6.1.2.1.25.3.3.1.2.2 for DP-System
• .1.3.6.1.2.1.25.3.3.1.2.3 for DP-Packet
In PAN-OS 4.0, these MIB objects were re-organized into two hrProcessorLoad OIDs:
• .1.3.6.1.2.1.25.3.3.1.2.1 for Management Processor
• .1.3.6.1.2.1.25.3.3.1.2.2 for Slot-1 Data Processor
This is from https://live.paloaltonetworks.com/servlet/JiveServlet/previewBody/4627-102-3-13273/SNMP_TechNote-RevA.pdf
It's inconvenient that we can't poll disk utilization, memory, etc yet, although I have heard that this is coming soon. I really wish it was as simple to monitor as other network devices, like the ASAs. There does not seem to be a way to get the standard Node Details page to include the CPU statistics gathered from a UnDP.
