zabbix faulty alarms about memory consumption on linux

[jamesit]

Hi all ,

I will bigin by discribing the setup , We have a Zabbix server instance installed on CentOS machine ,where we monitor all sort of servers from windows to linux ...

To monitor free memory on linux servers we are using the item 'free memory' that comes with the template 'Template OS Linux Active' with a small adjustment the item is changed to be

(vm.memory.size[available])

and the trigger used is as follow :

({TRIGGER.VALUE}=0 and {Template OS Linux Active:vm.memory.size[available].avg(1800)}>90 ) or ({TRIGGER.VALUE}=1 and {Template OS Linux Active:vm.memory.size[available].avg(1800)}<80 )

All good so far , except that we keep getting alarms about 'Free memory' all the time , but when i take a look at the graphs no trace of the problem

for example just yesterday we received this alarms many times , when the graph looks like this


When we get these alarms :
AT Tue 12/1/2015 12:28 PM Trigger: Amount of free RAM on the server Server for last 30 min less than 10% Status: OK Last value: 7.34 GB Trigger URL:

I checked the zabbix server logs , but with no success .
Have anyone any idea why do we get these false alarms , or for possible reasons ? any help is very well appreciated