Have successfully added CVM’s and ESXi hosts with SNMPv3 monitoring in Solarwinds but how about monitoring the AHV hosts? I have not really seen any guide for this.

@Tedevil

Monitoring of AHV hosts is established as part of cluster monitoring. There are hardware and host related messages triggered by the cluster (CVMs).

We have deployed Solarwinds as our global monitoring tool of ALL devices in our global network. So we really want monitoring of AHV hosts from that platform. It is our “all in one tool” for monitoring, notifications, statistics and reporting. I actually hoped that the SNMP configuration used in PE actually also pushed it out to the AHV hosts and not only the VM’s, but it did not.

Looking on the active services/processes on the AHV it seems SNMP is not used so I guess it would be pretty straight forward to enable this (install the net-snmp package). I realize that it is maybe not supported as it potentially could break things but perhaps something that can be added in the future?

@Tedevil

The monitoring of AHV hosts is performed by CVMs that will raise alerts and send SNMP traps. This is defined in the hypervisor section of the MIB. That is why SNMP configuration is not pushed to AHV hosts by PE.

Yes, but, snmpd is installed on the AHV hosts though it’s not adequately configured or systemctl enabled. 

SNMP monitoring of the hardware provides additional insights.  Sure it’s nice to have the CVM sending Nutanix specific traps but the hardware CPU stats (etc) are not true.   In our setup, the AHV hosts have 40 CPUs while the CVMs only have 10.  Graphing the CVMs is misleading.

I agree with @Tedevil and others.  Nutanix should fully support net-snmp on the physical hardware and provide a guide to setting it up.

Hello, I was wondering if anyone else was having issues where SNMP v3 isn’t returning a complete list of VMs using the Nutanix-MIB:vmName (OID:1.3.6.1.4.1.41263.10.1.3)

I have tried using both Solarwinds UnDP tools and SNMPWalk and the results of both seem to have omissions. 

Looking at the ROWIDs returned there are gaps and I wonder if these correspond to my missing vmNames.

.1.3.6.1.4.1.41263.10.1.3.13
.1.3.6.1.4.1.41263.10.1.3.14
.1.3.6.1.4.1.41263.10.1.3.15

.1.3.6.1.4.1.41263.10.1.3.18
.1.3.6.1.4.1.41263.10.1.3.19
.1.3.6.1.4.1.41263.10.1.3.20

As far as I can tell SNMP v3 is setup correctly. It certainly returns the correct number of Hypervisors along with their names for each of our clusters. 

Any help greatly appreciated, Andy

Indeed. I am disappointed by nutanix snmp monitoring abilities. I can live with this, but...