Fetch traces, inspect spans, and review experiments via the Phoenix CLI.
npx skills add https://github.com/Arize-ai/phoenix --skill phoenix-cli
Please help me install this Agent Skill. Skill: phoenix-cli Source: https://github.com/Arize-ai/phoenix/tree/main/skills/phoenix-cli Command: npx skills add https://github.com/Arize-ai/phoenix --skill phoenix-cli
The Phoenix CLI enables debugging and diagnostics of LLM-powered applications by fetching traces, inspecting spans, and reviewing experiments and datasets.
Install the Phoenix CLI globally and start debugging:
High-intent search queries and answers about installing and using this skill.
To debug LLM applications, you can fetch traces and inspect spans to identify failure points and latency bottlenecks in your interactions. The Phoenix CLI retrieves this data to provide quick insights during debugging sessions.
The best way to retrieve traces from an LLM project is by connecting the Phoenix CLI to your hosted project and executing commands like 'px traces' with configurable limits to fetch the specific trace data needed.
Yes, you can review experiments and datasets for LLM performance validation. The CLI fetches experiment results and dataset configurations to validate performance and ensure reproducibility across development environments.
You connect the Phoenix CLI to a hosted project by installing it globally via npm and running commands that target your Phoenix-hosted project, allowing you to retrieve traces, datasets, and experiment results directly.
Yes, the Phoenix CLI supports configurable limits for trace retrieval. You can execute commands with specific limit parameters, such as 'px traces --limit 10', to control the volume of data fetched during observability reviews.
You need observability for LLM-powered applications to diagnose errors and pinpoint the origins of issues in interactions. Fetching trace details allows software engineers to investigate latency bottlenecks and failure points across environments.
arsityTrace, evaluate, and monitor LLM applications with OpenTelemetry and Phoenix.
aRustyDevTrace requests across microservices using OpenTelemetry with Jaeger and Tempo.
qcmuuCapture OpenTelemetry traces and evaluation results for LLM debugging.
akillnessCoordinate LLM observability, tracing, and evaluation across multi-framework deployments.
