[bug] Ensure CLI report and json report use same data

[arekay-nv]

There is a discrepancy in the numbers reported by the CLI and in the reports generated to file. We should have the same results and be logging the events while generating a report at the end.

[anandhu-eng]

logging the events while generating a report at the end.

Currently, the errors are also being logged to the benchmarks_results.json file. What other events should we consider capturing? For example, we could include warnings, number of workers, scheduler type, or the adapter used by the endpoint client. These are some parameters I came across while reviewing the flow.