arize-traces

Name: arize-traces
Availability: InStock
Author: Steven

Analyze Arize AX traces to debug LLM issues and monitor trace health.

Updated Apr 7, 2026

Search Tags:#traces #token-usage #llm-observability #arize #ml-monitoring #trace-debugging

Ready to run

npx skills add https://github.com/stevenmunoz/turbo-ai-exercise --skill arize-traces

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: arize-traces
Source: https://github.com/stevenmunoz/turbo-ai-exercise/tree/main/.claude/skills/arize-traces
Command: npx skills add https://github.com/stevenmunoz/turbo-ai-exercise --skill arize-traces

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires certifi, and includes scripts (resource) components.

What problem does it solve?

Analyzes Arize AX traces to help ML teams diagnose issues, assess observability, and optimize model behavior by surfacing structured trace data, token usage, and health signals.

Core Features & Use Cases

Trace exploration: fetches and visualizes traces, showing hierarchical span relationships, timings, models, and status.
Health and cost insight: aggregates latency, error rates, and token usage for LLM spans to detect anomalies and cost spikes.
Use Case: Imagine you are debugging an failing LLM call; this skill helps you quickly locate the failing span, inspect inputs/outputs, and trace the error to its source.

Quick Start

Run python3 scripts/get_trace.py <trace_id> to visualize the full trace tree for a given trace.

Frequently Asked Questions about arize-traces

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I debug LLM traces to find a failing span?▼

To debug LLM traces, you can fetch and visualize the hierarchical trace tree to quickly locate the failing span, inspect its inputs and outputs, and trace the error to its source.

How do I monitor LLM token usage and trace health for anomalies?▼

You monitor LLM trace health by aggregating latency, error rates, and token usage across spans to detect anomalies and identify cost spikes in your model behavior.

Does this skill work with Arize AX for LLM observability?▼

Yes, this skill analyzes Arize AX traces specifically for projects using Arize for LLM observability, applying to spans and traces to surface structured data and health signals.

What is the best way to visualize trace trees from the command line?▼

The best way to visualize trace trees via CLI is running the get_trace script with a specific trace ID, which displays hierarchical span relationships, timings, models, and status.

Do I need Python to run trace exploration scripts?▼

Yes, you need Python to run the CLI-based exploration scripts that fetch and visualize traces, with the certifi package required as a dependency for handling SSL certificates.