My environment has multiple vCenter servers, with each vCenter instances having 4-7 ESXi hypervisors, and each vCenter instance has about 50-100 guests.
I can detect all of the vms from my probes, and I can monitor them individually, but the when I add a number of them, then I start getting alarms. The monitoring is flaky and unreliable at best even for 15 guests.
I get better performance when I monitor at the ESXi hypervisor host level.
Is there any way to combat tweak this? My scanning interval is 60 seconds. I don't want to monitor less than that.
Hello,
I'm afraid we would really love to make those sensors have a better performance, but that's beyond our reach, the way they have to get their information in the moment (via VMwares SOAP API and the SDK). However there are a few things that you can do How can I increase the connection limit on VMware systems? PE121, next to of course using longer scanning intervals or if the PRTG Probe is running in a VM, try to assign as many CPU-Cores as possible to it then.
best regards.
Oct, 2012 - Permalink