skills.rest
ethan-rio
ethan-rioCOMMUNITY
Explore all skills by this author →
GitHub

arize-instrumentation

Add end-to-end tracing and observability to LLM applications via Arize AX.

Updated May 5, 2026
Search Tags:#langchain#observability#openai#opentelemetry#session tracking#llm tracing#arize
One-click install
npx skills add https://github.com/ethan-rio/skills --skill arize-instrumentation-ethan-rio
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: arize-instrumentation
Source: https://github.com/ethan-rio/skills/tree/main/skills/arize/arize-instrumentation
Command: npx skills add https://github.com/ethan-rio/skills --skill arize-instrumentation-ethan-rio

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

LLM applications lack built-in end-to-end tracing and observability, making it difficult to debug model errors, track performance across agent workflows, and audit tool execution for production use cases.

Core Features & Use Cases

  • Two-phase agent-assisted flow: Analyzes your codebase first to detect stack, providers, and frameworks before implementing instrumentation, with user confirmation before any code changes.
  • Multi-language and framework support: Works with Python, TypeScript/JavaScript, Go, and Java apps, with auto-instrumentation for all major LLM providers (OpenAI, Anthropic, Bedrock, etc.) and agent frameworks (LangChain, LlamaIndex, Vercel AI SDK, etc.).
  • Custom span support: Adds manual CHAIN and TOOL spans for custom agent loops and tool execution to capture full input/output data in traces, avoiding sparse trace data.
  • Secure credential handling: Guides users to use the Arize ax CLI for profile management and never embeds raw secrets in generated code, referencing environment variables instead.
  • Use case example: If you have a Python LangChain agent that calls OpenAI tools to answer user queries, this skill will automatically instrument the LLM calls and add manual spans for each tool invocation so you can see full request/response data in Arize AX.

Quick Start

Use the arize-instrumentation skill to add full end-to-end tracing to your LLM application and start monitoring model calls and agent behavior in Arize AX within minutes.

Frequently Asked Questions about arize-instrumentation

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

FAQPage Schema
How do I add LLM observability and end-to-end tracing to my LangChain application?▼

You can add LLM observability by using Arize AX instrumentation to automatically trace LangChain workflows and LLM provider calls. The skill analyzes your codebase and applies auto-instrumentation to capture full request and response data for debugging.

Does OpenTelemetry-compatible tracing support Python, TypeScript, Go, and Java LLM apps?▼

Yes, OpenTelemetry-compatible tracing supports Python, TypeScript/JavaScript, Go, and Java LLM applications. It provides auto-instrumentation for major providers like OpenAI and Anthropic, alongside agent frameworks such as LlamaIndex and Vercel AI SDK.

How do I create manual spans for custom agent loops and tool execution?▼

You create manual spans for custom agent loops by adding specific CHAIN and TOOL spans to your code. This captures full input and output data in traces, preventing sparse trace data when monitoring custom tool execution workflows.

Can I use auto-instrumentation for OpenAI calls without exposing API keys in my code?▼

Yes, auto-instrumentation handles secure credential management by referencing environment variables instead of embedding raw secrets in generated code. It guides you to use the Arize ax CLI for profile management to keep API keys secure.

What is the best way to debug model errors and track performance across agent workflows?▼

The best way to debug model errors is implementing end-to-end LLM tracing with Arize AX to eliminate blind spots. It monitors agent behavior and audits tool execution by capturing full request and response data across your application.

Why are my LLM traces showing sparse data for custom tool execution loops?▼

LLM traces show sparse data when auto-instrumentation does not cover custom logic. You need to add manual CHAIN and TOOL spans for your custom agent loops and tool execution to capture complete input and output data in Arize AX.

Related Skills

Explore More Software Engineering Skills→
ethan-rioethan-rio

arize-ai-provider-integration

Manage LLM provider credentials as Arize AI integrations via the ax CLI.

Community
Intermediate
stevenmunozstevenmunoz

arize-traces

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

Community
Intermediate
PostHogPostHog
58

instrument-llm-analytics

Instrument LLM calls to emit $ai_generation events to PostHog.

Official
Intermediate
kagentikagenti
292

genai:semantic-conventions

Standardize AI agent instrumentation with OpenTelemetry semantic conventions for LLM observability.

Official
Intermediate