-
Notifications
You must be signed in to change notification settings - Fork 128
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
About CRI metrics of virtlet-based pod. #723
Comments
@jellonek Any suggestion? |
We don't have at the moment this topic on our agenda, but contributions are welcomed. CPU stats could be probably gathered from libvirt in similar manner to https://libvirt.org/docs/libvirt-appdev-guide-python/en-US/html/libvirt_application_development_guide_using_python-Guest_Domains-Monitoring-vCPU.html As for image, imo we should return only Allocation value of root disk. |
Ok, we have that on our agenda (i missed it) and probably in near future i'll start working on that. |
Filesystem stats were implemented in #735 Can you verify if that works for you, by testing |
@jellonek OK, Thanks for this feature. :-) |
All of that is already merged and included in last release. |
@MiaoY please note that when Virtlet is used together with docker you still will not be able to get VMs stats using Kubelet Summary API. It's because how kubelet is designed. If you want to get Virtlet VMs stats using Summary API then you need to switch to containerd. You can also just directly call ContainerStats then it will work everywhere. |
Directly - means on the node using CRI call on virtlet unix socket. |
Now, k8s mainly uses cAdvisor to collect container performance metrics, and the runtimes that are not included in the cAdvisor, if they have implemented
ContainerStats
of CRI, also can collect some of metrics.At present, we need to autoscale Virtlet Pods according to the resource usage of VM. However, virtlet has not yet implemented the
ContainerStats
of CRI, and there is no Virtlet Pod metrics data collection in cAdvisor also.Are we going to implement the CRI metrics in the near future?
At present, how can I get the ContainerStats(VM metrics) of Virtlet-Based Pod?
In addition, k8s achieves the Images garbage collection by collecting the disk usage of the container/image. Is there any way to collect the Virtlet image?
Hope to get someone's help, thank you very much!
The text was updated successfully, but these errors were encountered: