LVE Stats. How it works

First of all, there are three processes lvestats-server.py: 

root 24006 0.1 1.7 312132 32620 ? SN Nov28 1:51 /opt/cloudlinux/venv/bin/python3 /usr/share/lve-stats/lvestats-server.py start --pidfile /var/run/lvestats.pid
root 24007 0.4 1.7 459596 32364 ? SNl Nov28 6:10 /opt/cloudlinux/venv/bin/python3 /usr/share/lve-stats/lvestats-server.py start --pidfile /var/run/lvestats.pid
root 24030 0.2 2.4 329940 45252 ? SN Nov28 3:45 /opt/cloudlinux/venv/bin/python3 /usr/share/lve-stats/lvestats-server.py start --pidfile /var/run/lvestats.pid

This is normal behavior, these processes are doing different things.

The main process handles the stats parsing routine, etc, while two others are working with the plugins.

The main lvestats-server.py process parses the /proc/lve/list file every 5 seconds. 

The time interval could vary, it depends on the plugins installed, but usually, the average query time is around 5 seconds.

Next, the gathered information is stored in the local SQLite database, but the locating can be easily changed via lve-stats configuration file. 

Lve-stats can store the data in MySQL (MariaDB) or PostgreSQL servers.

Next, the end-user or server administrator can select a desired time period and check the amount of the fault (if there were any) for all users or a particular user.

The default average value (A column in the graph) is 1 minute:

Column L shows the current limit, and F shows the number of faults.

The chart can be generated from the command line as well.

And finally, the end user can see the processes snapshot directly from his panel, for example: 

The snapshots are available for admin from the command line as well.