Volume Disk Space Utilization in Large Environment with different Threshold Values - Simplified

Version 1

    Hi Everyone,

     

    This document is to manage volume disk space utilization alert in a large environment with varying threshold value for different sized volume disks. It is not fair to set 90% disk utilization threshold for 10 GB, 100 GB and 10 TB or more volume drives, instead we need to have such a logic where we can define threshold values based on the different sized volume drives.

    I don't know whether some one already have done this before or anything like-  I had done. Here we go - We need two alerts to cover all types volumes for varying threshold.

     

    Alert 1

    First off you need to have volume data in excel sheet, you can directly pull that information from solar winds if you are monitoring volumes with nodes using export custom properties, or simply create a report for volumes with below fields-

    Volume Name, Volume Type, Volume Size, Used Space, Free Space along with node details like name and IP address. It would be better to pull average of last 1 week for used space field for understanding the speed of space utilization.

    After this categorize the volumes with in certain size limits for example -

    Cat A - 1KB to 1 MB

    Cat B - 1 MB to 50 MB

    Cat C - 50 MB to 250 MB

    Cat D - 150 MB to 500 MB

    Cat E - 500 MB to 2 GB

    Cat F - 2 GB to 20 GB

    Cat G - 20 GB to 50 GB

    Cat H - 50 GB to 250 GB

    Cat I - 250 GB to 500 GB

    Cat J - greater than 500 GB

     

    The count of categories are too much, well it depends on the volume data that you have in your environment also you can ,merge or modify the volume categories based on the average values like merge Cat - C and D.

    Now decide the threshold for those categories like below-

     

    Cat A - I will not monitor those volumes since they will be always 100 used. Or will put 100% as the threshold.

    Cat B - 99%

    Cat C - 98%

    Cat D - 96%

    Cat E - 95%

    Cat F - 90% etc.......

    Deciding threshold again depends on the input from the windows team and the application owner using the server, so there we need a good understanding between the different teams and coordination.

     

    now create alert using above conditions with OR logic for each Category. Like

    "Volume Space greater than 1 KB" AND "Volume Space less than 1 MB" AND "Volume Percent used is 100%" AND "Volume Custom Threshold custom property is empty"

    OR

    "Volume Space greater than 1 MB" AND "Volume Space less than 50 MB" AND "Volume Percent used is 99%" AND "Volume Custom Threshold custom property is empty"

    OR

     

    OR and so on

    Note : Volume Custom Threshold custom property will be explained in Alert 2.

    This alert will cover all the volumes types - like Mount Point, Virtual Disks, Physical Disks etc, so you can put conditions to ignore specific type of disk types.

     

    Alert 2

    Though most of the volumes will be covered using above method still if we have any exceptional case where volume threshold is extremely different than the standard values then we will go for custom threshold alert mechanism.

    First we have to create a volume custom property - I would name it as "Custom_Threshold" and data will be Integer value for it or any other format to support threshold values.

    You can put the threshold values in it for a exceptional volume like there is 20 GB volume drive where you need an alert when 500 MB is left i.e. 97.5% then simply put 97.5 in it.

    Now create another alert - I would name it as Custom value Alert and put conditions like below-

     

    Above I have used "Custom Threshold" custom property flag to identify which volume needs to alert on custom threshold value instead of default alert, we can use conditions like -

     

    "Volume Custom Threshold custom property is empty" AND " Volume {Percent Used} greater than or equal to {Custom Threshold Value}"

     

     

    Note: This method is little effort putting like-  to collect data of your environment and need a lot of analysis, but once you are done with this then you need to group volumes based on nodes, names or other parameter. just these two alerts will cover everything you need.

    Hope this will help.

     

    Thank you

    Nikhil Kumar Sharma