Viewing Statistics
Selectel Cloud Platform users can view statistics on the workloads of their servers. All data is presented in the form of graphs.
To view servers workload data in the Control panel:
- Go to the Servers section.
- Open the server page.
- Go to the Statistics tab.
The following metrics are available for each server:
- CPU usage;
- memory usage;
- volume workloads;
- network interface workloads.
Statistics for all metrics are also available through the REST API. More information can be found in our blog in the VPC Statistics article.
VirtIO Drivers
VirtIO drivers are paravirtualized drivers for guest operating systems running in a KVM. They provide direct (paravirtualized) access to virtual machine devices and equipment to avoid slow emulated connections. VirtIO drivers are used for gathering statistics and, in some cases, the guest machines require additional setup.
Please note that the metrics are not collected for disabled machines.
Delays
Metric values of virtual machines are gathered every five minutes.
Please note that it takes ten minutes to generate the first set of values for network interfaces and volumes. This is because the difference between the first two sets of measurements (at the five- and ten-minute mark) has to be calculated for these metrics.
Linux Systems
The required drivers are available in all modern Linux distros by default. However, if no metrics are gathered for your server, you can add an auto-load kernel module:
- for Debian-based distros:
Add the following lines in the /etc/modules file:
virtio_blk
virtio_scsi
virtio_balloon
virtio_net
- for RedHat-based distros:
Add the following lines in the /etc/rc.modules file:
modprobe virtio
modprobe virtio_net
modprobe virtio_balloon
modprobe virtio_scsi
modprobe virtio_blk
These modules will automatically launch when the server restarts.
If you cannot restart your server, you can run the following commands in the terminal:
modprobe virtio
modprobe virtio_net
modprobe virtio_balloon
modprobe virtio_scsi
modprobe virtio_blk