Hello,
I know that Predictive Monitoring exists for CPU, Hard drive and Interfaces ( Predictive monitoring ). In our company there are certain VMs whose RAM load follows a very predictable pattern (See Screenshot below).
It would be beneficial for us if the usage pattern could be detected and to avoid having a CRIT being pushed to the dashboards if it’s just a regular usage pattern.
Is it possible to have predictive monitoring active for the RAM usage of certain VMs?
Best regards
As far as I know this feature (Predictive Monitoring for Memory) is only supported for Windows machines. To begin you need to understand how you will base your prediction (Day of the week, day of the month, hour, day or minute) then based on that you will allow Checkmk to define what is expected for that period and any deviation can considered as an alert:
From my perspective the predictive monitoring is more useful to identify deviations on really stable environment where a deviation on memory from 20% to 70% from one week to another needs to be investigated. One option to reduce the noise, if that is the objective, is to enable the option for Averaging which will avoid alerting during peak utilization.
