Our organization has a lot of services published to use shared instances due to their usage being minimal. Since there are so many of these services on our servers it can be hard to get a full picture of how our available shared instances are working. We want to better understand how shared instances are being used and how it effects our server’s resources.
Currently in ArcGIS Monitor and ArcGIS Server there is no way tell how many total shared instances are being used or any other metrics. I would like to be able to track total requests, instance saturation percent, and request time for shared instances to better understand and calibrate our server.
Our 2 servers currently have it setup so there are 16 shared instances per machine and a max of 50 cached services per shared instance. I am not sure if having this many is necessary as we have needed to increase it over time due to seeing lots of DynamicMappingServer log errors. We do still see lots of DynamicMappingServer log warnings, so maybe this could help troubleshoot these as well.
The only thing we really are able to do right now is get metrics on individual services that are using the shared instance pool, perhaps there is something we can do with that but I am not sure. Not quite sure what instance saturation with individual services mean. I looked at the DynamicMappingService with ESRI support, but this service does not appear to provide measurable metrics in ArcGIS Monitor or ArcGIS Server at least in regards to the use of shared instance services(Ex: Total Requests does not change if you use a shared instance service).
Here is a community post with a similar question posed: https://community.esri.com/t5/arcgis-enterprise-questions/monitoring-arcgis-server-shared-services/m...
