the value for load average seems to be set to 3 disregarding the cpu count. Is it possible to increase this value to say 12?
additionally I'm looking for way to send "CPU Overload" Email only if the CPU load lasts more then 5 minutes.
You cant increase the 5 minute load average incident you get in the dashboard I believe..
As far as the email load notification, that is easy enough, here you go ;-)
The attached ruleset will increment a counter every minute where the CPU is high. If it is not high anymore it will reset the counter. For this example, after 10 increments (10 minutes) you can configure it to send a notification.
you can change the default value for load average threshold in mwg-monitor configuration file.
Create the file /etc/mwg-config/mwg-monitor.conf and add a line "THRESH_LOAD=12.0".
# mkdir -p /etc/mwg-config
# echo "THRESH_LOAD=12.0" >> /etc/mwg-config/mwg-monitor.conf
# service mwg-monitor restart
thank you for this information, I've followed these steps and I'll check tomorrow if it helps.
BTW I see the number of cpu cores (ncpu) has been already used to calculate load_5m value, so I was wrong saying "... disregarding the cpu count".
I see these messages only on virtual appliances with two cpu cores, so may be I need to add some CPUs.