Add monitoring hooks and harden Triton runtime setup

[ryanznie]

Summary

• add Prometheus monitoring hooks and a endpoint
• add monitoring metrics collection for heuristic, model, and fallback paths
• harden LayoutLMv3 processor loading with local compatibility fallbacks
• add runtime support for LayoutLMv3 image processing
• add monitoring and traffic-generation documentation

Testing

• python3 -m py_compile app.py src/api.py src/inference.py
• python3 -m py_compile scripts/generate_traffic.py
• exercised Docker/Triton startup iteratively during integration

[github-actions]

Coverage report

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
.
app.py
Project Total

This report was generated by python-coverage-comment-action