Database workloads may show different resource usages for database applications. Database administrators can enhance the DBMS performances through resource management that reflects workload characteristics. We provide a method that can identify tunable resources from analyzing the relationship between performance indicators and resources. First, we select which performance indicators increase or decrease by expanding resources using a correlation coefficient and a significance level test. Next, we identify resources that can affect the DBMS Performances by using increasing or decreasing performance indicators. We evaluated our method in the TPC-C and TPC-W environments.