Observability & Evaluation · Data artifact

Failure Signature Catalog

Data artifactObservability & EvaluationObservability & Evaluationarc:FailureSignatureCatalog

A curated playbook mapping characteristic trace patterns to failure modes (tool selection, parameter generation, interpretation, state, hallucination, reasoning errors) and their root causes and fixes.

Responsibility. Records known trace signatures of failure modes to guide classification and targeted verification.

Also known as: Debugging playbook, Error pattern catalog, Logical error taxonomy, Coordination failure signatures

Direct neighbourhood (hover for relationship types)

Relationships

configures structural

Reasoning Verifier abstract Ch3.6
Trace Error Classifier Ch3.6

is read by dependency

is produced by lifecycle

Failure Analyzer Ch3.6

Design guidance

SHOULD be maintained from recurring failures surfaced in trace reviews and used to focus verification on common error types.

Quantitative guidance

As stated by the sources; verify before use.

Example: 30% of failures attributable to scope confusion (Ch3.6).
Deadlock signature: simultaneous mutual dependency waits with no child spans until 45 s timeout; out-of-order signature: dependent agent starts with dependencies.count = 0 before producer completes; state signature: version_conflict events (Ch8.2A).

Classification

Patterns: Error pattern extraction
Quality attributes: Maintainability (ISO/IEC 25010)
Risks mitigated: Recurring one-off patches instead of systematic fixes

Sources

Ch3.6: T. Nguyen, "Trace Analysis and Execution Debugging," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 3.6. ISBN: 9798244538229.
Ch8.2A: T. Nguyen, "Error Rates and Reliability," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 8.2A. ISBN: 9798244538229.
Ch10.4: T. Nguyen, "Human-in-the-Loop," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.4. ISBN: 9798244538229.