We have a web application that consist of many components. Lets say I have a webserver01 with sql cluster (sql01, sql02, sql03). I want to create a availability meter for customers to see. How should I configure out Site24x7 to handle these following cases:
Webserver goes down and sql cluster is untouched (all green). Application is down and should decrease availability.
sql01 goes down. Sql02, sql03 and webserver01 is up. Cluster health is yellow (cluster is not in ideal state) but actual requests goes through. This should not decrease availability since application is working correctly.
sql01-03 goes down. Webserver01 can't make any database queries and thus the application fails. This should affect availability.
I've tried to play with monitor groups, subgroups and thresholds but to me it seems the relationships are the wrong way round. I've made sql cluster as a subgroup and webserver01 as a main group. If subgroup is down the webserver01 main group should also go down but it stays green. Is there something I'm missing here or how would you suggest to build such availability monitoring system (if possible)?