Observability & Evaluation · Data artifact
Performance Baseline
Data artifactObservability & EvaluationObservability & Evaluationarc:PerformanceBaseline
A recorded reference of latency percentiles (p50/p95/p99), request and token throughput and GPU utilization under representative load and default configuration.
Responsibility. Provides the before-state against which optimizations and regressions are judged.
Also known as: Baseline measurement, Historical performance baseline
Relationships
is read by dependency
is written by dependency
is produced by lifecycle
Design guidance
- MUST establish a baseline before optimizing to quantify improvement and set per-operation performance budgets.
- SHOULD be captured before schema or configuration changes and compared after deployment to detect regressions.
- MUST record the concurrent load alongside each latency percentile (e.g., 'P95 1.2 s at 100 RPS').
- SHOULD be recaptured after each optimization to validate predicted impact through a before/after comparison.
Quantitative guidance
As stated by the sources; verify before use.
- Interactive applications target p99 latency under 2 seconds; enterprise scale targets 1000+ requests/second (Ch4.4).
- Before/after tool-result caching over 10,000 traces: mean 1,300->1,075 ms (17%), P50 1,100->900 ms (18%), P95 3,200->1,800 ms (44%), P99 7,500->3,100 ms (59%) (Ch8.1).
- Baseline LLM inference ~600 ms vs 1,800 ms during a GPU memory incident (Ch8.1).
Classification
- Quality attributes
- Performance efficiency (ISO/IEC 25010)
- Risks mitigated
- Confusing measurement variance with improvementPremature optimization when targets are already met
Sources
- Ch4.4: T. Nguyen, "Performance Profiling and Optimization," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 4.4. ISBN: 9798244538229.
- Ch6.2B: T. Nguyen, "Production Vector Database Deployment," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 6.2B. ISBN: 9798244538229.
- Ch7.3: T. Nguyen, "NeMo Agent Toolkit Profiling," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 7.3. ISBN: 9798244538229.
- Ch8.1: T. Nguyen, "Latency Metrics," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 8.1. ISBN: 9798244538229.