Community
Select to view content in your preferred language

Need help understanding the Average Instance Used Metric.

145
1
3 weeks ago
MichaelBartlett
by
Regular Contributor
3 weeks ago

The Avg_Instance used metric seems to be pretty low, rarely over 1, even though in Sever Manger I will often see the Instances used jump up to 8 instances are more.  The Instance saturation is well over 50%.

I am assuming the actual time it is in use is very small, hence the small average.

So what would be a good way of judging what these numbers mean?  

I would think some number range that corresponded closer to the Instance in Use data shown in Server Manager would be more useful for tuning and setting the Min and max Instances on a service.  Is there a different metric for it? (Max Instances in Use??)

Configure service instance settings—ArcGIS Server | Documentation for ArcGIS Enterprise

"This chart is another good way to identify usage trends on a per-service basis and can help answer questions such as “What ArcGIS Server services might be good candidates to move to the shared instances pool?” If a service uses few instances on average, you may want to move it to the shared pool. If a service uses many instances on average, it should probably have a set of dedicated instances to process incoming requests."

Appreciate any guidance you can give.

 

1 Reply
StevenBeothy
by
Regular Contributor
a week ago

Hi Michael
Are you still on ArcGIS Monitor 2023.x? This is the older release where I see an Instances Used Avg metric, post 2024.x I now only see Instances Used with a good description of how this metric is calculated.

Both Instances Saturation Percent (%) & Instances Used Avg are available at the GIS Server site level:

StevenBeothy_1-1761666203261.png

StevenBeothy_0-1761666141209.png

and GIS Service levels. I believe the metrics you are most interested is the Service level. In my example below the USDM_2019_Drought service is within the Shared Instance pool with 4 min and max instances. When this service receives a high number of requests (30 req/s) it saturates the shared instance pool to 49% with an Avg Instances in Use of 2.

StevenBeothy_3-1761666310227.png

StevenBeothy_2-1761666274585.png

StevenBeothy_4-1761666396921.png

I would recommend to read this very good Blog posting by Derek Law titled ArcGIS Monitor: Understanding Component Status and Alerts.

If you are looking for an excellent and easy to use utility with your ArcGIS Monitor metrics look into the ArcSOC Optimizer for ArcGIS Monitor ,this utility can inform you on which services need to be moved from Shared to Dedicated instances (or vice versa) and how to configure your Min and Max instances.

ArcSOCOptimizer is a python command line utility.  It can help administrators:

  1. Reduce memory by identifying low usage services with dedicated pooling and reducing the overall number of arcsoc.exe processes:
    1. changing dedicated to shared instance pool, if supported  
    2. reducing min instances for dedicated instance pool services
  2. Improve performance by identifying high usage service and allocating dedicated arcsoc.exe processes if available memory:
    1. changing shared to dedicated instance pool
    2. increasing min instances for dedicated instance pool 


Finally if you are on an older release of ArcGIS Monitor 2023.x please keep in mind the ArcGIS Monitor Life Cycle. Which shows ArcGIS Monitor 2023.3.1 being retired on August 27, 2025.