Linux on the mainframe provides an excellent opportunity for hosting many mixed workloads due to its balanced design of CPU, memory, and I/O bandwidth. The design balance will help you meet the expectations of the end users of these differing workloads, as codified in the SLAs. Should you wish to tune, find anomalies in application behavior, or plan capacity for the future, you can do so relatively easily.
In a Linux-on-the-mainframe environment using z/VM, abundant data are available for analysis. z/VM also provides tools for tuning. There are tools for gathering and analyzing data.
Since z/VM functions as the means to adjust guest resources and is a good data gatherer and analyzer, in many cases no further interpretation of the data is needed. The collected data are in a format that directly corresponds to the means of resource adjustment. While it is possible to build an entire performance management scheme entirely out of the z/VM environment, there is no need to limit yourself to that approach. There are many tools available from both software vendors and the Open Source community, so that any company can feel at home in doing Linux-on-the-mainframe performance management.
The mainframe has a different architecture and design from most other computer systems. Its resources, CPU, memory hierarchy, and I/O are primarily designed for resource sharing in mixed workload environments.