IBM Cloudability

Cloudability

A place for Cloudability product users to learn, connect, share and grow together.


#Cloudability
#Automation
 View Only
Expand all | Collapse all

Follow up on Elasticity Score supporting graph

  • 1.  Follow up on Elasticity Score supporting graph

    Posted Wed July 14, 2021 02:20 PM
    I want to thank @Dustin Bowling for responding to an earlier post about graphs representing Elasticity:

    https://community.apptio.com/communities/community-home/digestviewer/viewthread?GroupId=1639&MessageKey=f9e94648-a682-485c-8828-10731a7d07c4&CommunityKey=f67c7e7c-be1c-4053-9845-2376da697342&tab=digestviewer

    I want to follow up on this because our users use the Cloudability Scorecard feature to look up their individual application scores based upon business mappings that we have.  I can't come up with a graph that supports a score of '90' vs. a score of '20'.  I have attached two screenshots to demonstrate what we are seeing. One is a graph representing an application with a elasticity score of 25; the other is a score of 91. I have the dimensions of utilization hours and CPU Average.    I am asking for a little more insight as to what parameters the scorecard for Elasticity is measuring and hopefully I can show our users some supporting information for their scores. ​

    ------------------------------
    Ray
    ------------------------------

    #Cloudability


  • 2.  RE: Follow up on Elasticity Score supporting graph

    Posted Mon July 26, 2021 10:19 AM
    Hey Ray!

    My apologies for the delayed response and I appreciate the inquiry.  We have a couple of things at play here that we need to consider, namely 1) Actual usage of EC2 Compute by hour of day (chart provided) and 2) Cloudability Scorecard Elasticity metric which is strictly based on "autoscale" recommendations from the Cloudability Rightsizing engine.  The Usage chart is looking at what is running vs. what isn't, and the Elasticity score is looking at how well the stuff that is running is scaling up or down to meet workload demand.

    To further elaborate, the Elasticity score is looking at all of the compute resources where you can save on cost by autoscaling the resources. It is essentially doing a comparison of how much you can save vs. your total compute spend.  Therefore, if there are no recommendations of autoscale and you cannot save on any recommendations, then this score is 100, as the idea is that you are fully optimized if there is no further recommendation to optimize this. If you have a number of resources that you can save on, then this score decreases.  Note that this does not include spot, and will include the resources that you see on the Rightsizing page.  This score doesn't take into account CPU, Memory or Disk only at Cost data.

    I may have mislabeled the chart as it is more geared toward compute workload persistence vs. elasticity - my apologies.  Hopefully this helps clear up any confusion.

    Cheers!

    ------------------------------
    ---------------------------
    Dusty Bowling
    FinOps Principal
    Apptio
    ---------------------------
    ------------------------------