Overhead of collecting server monitoring metrics

Stardog provides a broad range of server monitoring metrics (e.g., Server Monitoring | Stardog Documentation Latest)

We think that collecting these metrics may impose some amount of overhead. We can disable monitoring completely by setting metrics.enabled to false in the stardog.properties. Based on this, we have a question. If we do so, these metrics are not gonna be collected at all or they're gonna be collected anyway but since the metrics.enabled is false, they're not gonna be printed out on the disk?

Thank you.