Issue Health Monitor on Onyx wrong memory

[zoni]

Hello,

On Onyx the health monitor is displaying memory 15% usage while on Tools and settings => server information is about 35% usage.

Is this a bug? How do I restart the HM service on centos 7?

Thanks.

 

[UFHH01]

Hi zoni,

you are able to restart the depending services with:

/etc/init.d/psa-health-monitor-notificationd restart
/etc/init.d/sw-collectd restart

Pls. see as well: => Plesk for Linux services logs and configuration files ( KB - article: 111 283 )

 

[zoni]

Hi,
I dont have that files inside /etc/init.d/ on Plesk Onyx with Centos 7

I have removed HM and installed it again but it still shows erroneous memory available, always arround 15%, when in reality is 40%, however drive space an cpu seems correct.

The app for android also shows the correct memory just as in system information so the problem should be with HM itself.

 

[UFHH01]

I dont have that files inside /etc/init.d/ on Plesk Onyx with Centos 7

If you use an operating system, without init - scripts, you would certainly use "systemctl" instead.

systemctl stop psa-health-monitor-notificationd.service
systemctl start psa-health-monitor-notificationd.service
systemctl restart psa-health-monitor-notificationd.service

systemctl stop sw-collectd.service
systemctl start sw-collectd.service
systemctl restart sw-collectd.service

OR

service psa-health-monitor-notificationd stop
service psa-health-monitor-notificationd start
service psa-health-monitor-notificationd restart

service sw-collectd stop
service sw-collectd start
service sw-collectd restart

​

 

[Bitpalast]

The difference in values can be due to different times of measurement. The server information data is collected at a different time than the health monitor data. Health monitor is an extension that works independently from the built-in server information page. It is thinkable that health monitor displays different CPU and RAM values, because these values can change frequently and within milliseconds.

 

[pacquo]

I am observing the same behavior after the upgrade of Plesk from 12.5 to Onyx on Ubuntu 14.04.5 LTS.

The difference between the real memory usage shown by the Plesk health monitor now appears constantly as:
the amount of [used memory minus buffers minus cache] minus [MySQL memory usage] minus [Plesk memory usage].

In other words:
the value of [Plesk memory usage plus MySQL memory usage plus real memory usage] shown by the Plesk health monitor, is now equal to the value of [used memory minus buffers minus cache) returned by the linux command free.

Before the upgrade the value of the real memory usage was, alone, equal to the value of [used memory minus buffers minus cache].

 

[zoni]

Here are some screenshots

HM:


Server information:



I think this is most likely a bug.