Monitor VM's resources (CPU, RAM etc)
Solution 1
There's a ganglia plugin for sFlow which is supposed to be able to capture this kind of information, but I've never tried it.
Solution 2
CollectD has a libvirt plugin that can query performance information of VM guests from the host. We are using this quite successfully to monitor OpenStack VMs.
http://libvirt.org/apps.html#monitoring has a few more tips, but I've only had experience with the CollectD plugin. The Nagios-virt package looks interesting and might be nice to try, but I just use a regular Nagios setup (i.e active, passive or NRPE service checks) for alerting with respect to VMs.
Related videos on Youtube
Simon
Updated on September 18, 2022Comments
-
Simon almost 2 years
I'm testing Openstack and i'd like to know how can i find out about resource utilization inside instances (Virtual Machines).
Openstack uses kvm, qemu and libvirt in my testing environment. Is there any way to get those stats without logging in machines (CPU, Memory, Swap, Network traffic)? I'd like collect those data from hypervisors or through libvirt.
Any suggestions?
-
Simon about 12 yearsI've read that libvirt plugin can't gather all stats, to get them all collectD has to be installed on guest OS and i'd like to avoid it, but despite this I'll try it out. 1 point up, but i'm still looking for software around libvirt, mayby there are other options.
-
Simon about 12 yearssFlow is supported only in trunk so it's not what i'm looking for, but ganglia itself is really nice software.
-
Simon about 12 yearsRight now I'm using simple libvirt stats. It gives CPU, network bytes In/Out and disk IO. If I need to extend it I'll use collectd.
-
Simon over 11 yearsfinally I decided to use ganglia. It was a bit difficult to set up everything and hook ganglia with host sFlow and with Openstack, but finally I was able to create useful PoC. Best in sFlow is that it supports many supervisors, including Hyper-V :)