Add support to not use dummy benchmark files for workers

[davidkel]

Currently the workers use dummy benchmark files because the workload information comes from the manager. But monitors such as pushgateway contain information the workers need and the worker is hardcoded to look in the benchmark file for this information to know if it needs to run a monitor so until things improve there, we should support workers using a specific benchmark file to enable monitoring.

We need to really improve this is in caliper, the only monitor of use is for the worker to report performance information to prometheus. The logging monitor is of little value and has problems as well.