generic service usage metrics

I'm trying to piece together how the usage metrics part of a system based generic service works

My generic service is based on a system which contains a 4 node DB RAC.

at the database instance level we have our standard monitoring metrics with their thresholds.

When I add usage metrics within the generic service I can also base them on database related metrics - a slightly different choice here than at the database instance metrics level

I then add my choices.

My question is is the generic service "sucking" up metrics from the database instance ?

Do I have to have the metrics enabled at the database instance level before I see values populated at the generic service level in the performance/incidents tab ?