Environment access + workload/data flow mapping
Model stress scenarios defined (latency, scale, edge cases)
Controlled tests across infra + monitoring performance
Benchmark scoring + root cause analysis for scaling/failure issues
Findings report, top 3 issues, and recommendations roadmap
