VM Compute report against previous month

Is it possible to generate a report on the changes in compute demands(CPU & RAM Usage) for each cluster. For example if CPU and RAM has changed for few VM's in the previous month and I'd like to compare with the current month which all VM's has increased the resources. Can anyone suggest if this is something feasible? using VMAN or SAM

  • VMAN stores historical resource usage of the individual VMs and of whole clusters, so you can do either way fairly simply using the GUI report builder and just creating two charts, one for the current time period and another for the previous month. There isn't a widget for directly overlaying two time spans over the top of each other.  You can get a lot more intricate with how you choose to display the data, but that is usually all going to boil down to writing custom SQL to pull it out of the database and massage the numbers however you think you want them to work.

    The big caution I will throw out there for anyone doing month over month reporting is your retention periods can be important.  Default behavior is to store detailed stats for 7 days, then hourly averaged data for a month, and any data older than 30 days is going to be daily averages. You lose a lot of detail as you move further back in time, i don't find daily averages to be operationally useful at all since I can't tell if something is 50% because it is medium loaded or because it was pegged out half the day and completely unused the other half.  Increasing retention is one solution, but it tends to make your database grow substantially and can slow down the loading times of some views.  There is a balancing act to all this, but as a general rule I try not to treat my monitoring tool as a long term data archive because full detail data over long periods of time just ends up being expensive to store and process.