We need to create a watch to calculate an accurate high memory alarm on Linux servers.
We need the following, unless there is a better way to monitor this please advise:
The alarm will be triggered when used Physical memory is more than 95% used and when the cache is a low actual value, such as 1GB or less. When these 2 are true then create a Critical.
We have never used a watch, We know We will have to mapped the necessary attributes to this equation and also make sure, we do not use this watch for all devices (we have a total of more than 7,000 devices , routers, switches, servers, etc). We were thinking of using a global collection to apply this watch to since it is only to 8 devices.
We have mapped the ucdavis attributes. Need help with the expressions syntax to get average.
Used memory / Total memory * 100 > 95% and memory cache < 1000 then when this is true create critical alarm
We are using our Dev environment to test this on.
Any help would greatly be appreciated.
Thank you