feat: Benchmark with more realistic data

[Leonardo-Ferreira]

The original benchmark for RBAC With Cache was a bit unrealistic because the same user and the same data was always requested, which is a highly cache-able scenario. By introducing a little variance on the the user and data actually enforced, we can see a MAJOR skew on large datasets but NOT on the smaller ones (this proves that the "randomization" is not the actual at-fault for the performance drop)

[casbin-bot]

@sagilio @xcaptain @huazhikui please review

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
3 out of 5 committers have signed the CLA.

:white_check_mark: hsluoyz
:white_check_mark: sagilio
:white_check_mark: Leonardo-Ferreira
:x: apiscevs
:x: dviry
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[sagilio]

It is a great point. But the old benchmark is synced from the go casbin repo, it is better to use a new file like EnforcerWithCacheRandomBenchmark.cs

[hsluoyz]

@Leonardo-Ferreira plz fix:

[Leonardo-Ferreira]

@hsluoyz this error happens because I cannot read the repos secret. If you look at the activity, it has successfully analyzed the code, but failed at the step of uploading the report due to a invalid key (it was not replaced). This issue is being fixed on a separate PR

[hsluoyz]

@sagilio