Feat/max error rate

[markVaykhansky]

Addressing this issue:
#105

--max-error-rate flag was added.
It is only applicable in either or the the following scenarios:

1. Fixed length dataset
2. max_requests is set
3. rate_type is constant and max_duration is set

• If error rate is not applicable (i.e non of the above scenarios) we just ignore it (throwing an error would break sweep for example)
• Implemented using multiprocessing.Event which is used to trigger a shutdown once we hit max_error_rate.
• max_error_rate - added to report
• error_rate in the final report is calculated as total errored / total requests finalized i.e we exclude incomplete requests

[nm-red-hat-upstream-automation-bot]

📦 Build Artifacts Available
The build artifacts (.whl and .tar.gz) have been successfully generated and are available for download: https://github.com/neuralmagic/guidellm/actions/runs/15167220831/artifacts/3170182036.
They will be retained for up to 30 days.

[nm-red-hat-upstream-automation-bot]

📦 Build Artifacts Available
The build artifacts (.whl and .tar.gz) have been successfully generated and are available for download: https://github.com/neuralmagic/guidellm/actions/runs/15167280131/artifacts/3170205369.
They will be retained for up to 30 days.