Provide the ability to set alert thresholds based on other parameters such as instance type, max instances and number of servers in the site. We have multiple sites with hundreds of services so setting the alert thresholds for each service is time consuming and hit and miss.
For example if we have an arcgis server site running 3 servers a service with max 10 instances per server we might say
threshold for req rate might be 10 req/s, per instance, per machine.
this would equate to 10 x 10 (instances) by 3 (servers)....the threshold would be 300 req/s.
the same site might have a service running max 3 instances per server
this would equate to 10 x 3 (instances) by 3 (servers)....the threshold would be 90 req/s.