Release Notes

List Traces by Project

March 13, 2026

Available in arize-phoenix 13.15.0+

A new GET /v1/projects/{project_identifier}/traces REST endpoint lists traces for a project with rich filtering, sorting, and pagination support. Use it to build custom trace browsers, feed traces into downstream pipelines, or retrieve traces by session.

# List the 50 most recent traces, sorted by start time
GET /v1/projects/my-project/traces?limit=50&sort=start_time&order=desc

# Include full span details inline
GET /v1/projects/my-project/traces?include_spans=true&limit=20

# Filter to traces from a specific session
GET /v1/projects/my-project/traces?session_identifier=my-session-id

• Sort by start_time or latency_ms in ascending or descending order
• Cursor-based pagination for consistent traversal of large trace sets
• include_spans=true to embed full span details per trace in a single response (batch-loaded to avoid N+1 queries)
• Session filtering via session_identifier — accepts plain session IDs or GlobalIDs; multiple values are OR-combined
• Time range filtering with start_time and end_time (ISO 8601)

Span Filters: Name, Kind, and Status Code

March 18, 2026

Available in arize-phoenix 13.15.0+ (server), arize-phoenix-client 2.1.0+ (Python), @arizeai/phoenix-client 6.5.1+ (TypeScript)

The spans API now supports filtering by span name, span kind, and status code. Pass any combination of these filters to narrow results without post-processing the full span list.

Filters are OR-combined within a field and AND-combined across fields — for example, name=["llm_call", "retriever"] & span_kind=LLM returns spans named llm_call or retriever that are also of kind LLM.

REST API:

# Filter by span kind
GET /v1/projects/my-project/spans?span_kind=LLM

# Filter by multiple names
GET /v1/projects/my-project/spans?name=llm_call&name=retriever

# Filter by status code
GET /v1/projects/my-project/spans?status_code=ERROR

Python SDK:

from phoenix.client import Client

client = Client()

# Get all LLM spans with errors
spans = client.spans.get_spans(
    project_identifier="my-project",
    span_kind="LLM",
    status_code="ERROR",
)

# Filter by multiple names
spans = client.spans.get_spans(
    project_identifier="my-project",
    name=["llm_call", "retriever"],
)

TypeScript SDK:

import { getSpans } from "@arizeai/phoenix-client/spans";

// Get all LLM spans with errors
const { spans } = await getSpans({
  project: { projectName: "my-project" },
  spanKind: "LLM",
  statusCode: "ERROR",
});

// Filter by multiple span names
const { spans: namedSpans } = await getSpans({
  project: { projectName: "my-project" },
  name: ["llm_call", "retriever"],
});