Skip to content
Arxo Arxo

Examples and Report Walkthrough

Examples and Report Walkthrough

Section titled “Examples and Report Walkthrough”

This page walks through real ml_architecture sample artifacts and how to interpret results.

Sample Artifacts

Section titled “Sample Artifacts”

Existing sample outputs:

  • crates/arxo-engine/src/metrics/ai_observability/ml_architecture/samples/fastapi-workflow-report.json
  • crates/arxo-engine/src/metrics/ai_observability/ml_architecture/samples/fastapi-workflow-report.md

Sample project and config:

  • crates/arxo-engine/src/metrics/ai_observability/ml_architecture/samples/fastapi-workflow/
  • crates/arxo-engine/src/metrics/ai_observability/ml_architecture/samples/ml-architecture-config.yaml

Regenerate the JSON Sample

Section titled “Regenerate the JSON Sample”

From your project directory, run Arxo with the path to your ML project and config:

Terminal window
arxo analyze \
  --path /path/to/your/ml-project \
  --config ml-architecture-config.yaml \
  --format json \
  --output report.json

How to Read the Report

Section titled “How to Read the Report”

1. Start with composite health

Section titled “1. Start with composite health”
  • ml_architecture.overall_score
  • ml_architecture.overall_score_extended

Use this as the top-level release-readiness signal before detector-level triage.

2. Triage by detector families

Section titled “2. Triage by detector families”
  • Core architecture: skew, boundary, pipeline complexity, reproducibility, lineage.
  • Evaluation and controls: eval integrity, data validation, CI integration, fairness.
  • Runtime operations: serving maturity, drift monitoring, monitoring/alerting, staleness, serving ops.
  • Rollout safety: A/B testing, shadow/canary.

3. Review diagnostics and evidence

Section titled “3. Review diagnostics and evidence”
  • ml_architecture.gpu_file_count, ml_architecture.database_file_count, ml_architecture.env_config_file_count
  • evidence-heavy modules and repeated low-score patterns
  • ml_architecture.graph.* entries for structure context

4. Prioritize fixes

Section titled “4. Prioritize fixes”
  1. Low-score categories in high-centrality modules.
  2. Controls with production blast radius: skew, lineage, eval, serving, drift.
  3. Rollout hardening: A/B, canary, alerting, staleness operations.

Suggested Interpretation Pattern

Section titled “Suggested Interpretation Pattern”
  1. If overall_score is low with weak skew + boundary scores, align train/serve interfaces first.
  2. If reproducibility/lineage remain low, prioritize immutable artifacts and dependency pinning.
  3. If serving/drift/alerting are low, harden operational controls before increasing rollout velocity.
  4. Re-run after fixes and verify upward movement in overall_score plus affected detectors.
Section titled “Read Next”