[Improvement] Benchmark Latency better Trial Identification

[gebauerm]

Currently Latencies of a benchmark are saved like this:

{
                "id": "pid_16579__name_deploy__objHash_8732029335278__id_0cc789b9-27f8-461a-ac4c-4919d7763d8a",
                "name": "deploy",
                "start_time": "2022-06-14 16:30:59.685459",
                "end_time": "2022-06-14 16:30:59.685464",
                "duration": 5e-06
            },

It can be hard for a user to identify a specific trial, as trials are by assumption encoded in the objHash which is not intuitive. The underlying assumption is that trainand validate of the objective are executed from the same object. If this is not the case anymore, trails can no longer be identified.

[gebauerm]

Metric Parent Class - which gathers basic system info and creates and ID