respan-instrumentation-chroma

Respan instrumentation plugin for Chroma Python applications.

The package adds Respan task spans for Chroma client and collection operations, including collection lifecycle calls, writes, reads, vector queries, updates, deletes, and counts. Embeddings and image payloads are summarized by count and dimension instead of being written as raw span attributes.

Configuration

1. Install

pip install respan-ai respan-instrumentation-chroma chromadb

2. Set Environment Variables

Quickstart

import chromadb

from respan import Respan, workflow
from respan_instrumentation_chroma import ChromaInstrumentor

respan = Respan(instrumentations=[ChromaInstrumentor()])


@workflow(name="chroma_query_workflow")
def run_chroma_query() -> dict:
    client = chromadb.Client()
    collection = client.get_or_create_collection("respan_docs")
    collection.upsert(
        ids=["doc-1"],
        embeddings=[[0.1, 0.2, 0.3]],
        documents=["Respan traces Chroma operations."],
        metadatas=[{"topic": "observability"}],
    )
    return collection.query(
        query_embeddings=[[0.1, 0.2, 0.3]],
        n_results=1,
        include=["documents", "metadatas", "distances"],
    )


print(run_chroma_query())
respan.flush()
respan.shutdown()

Further Reading

See the Respan example projects for runnable Chroma scripts.