Haystack
Get Enterprise Support
Get Started
🆕 Haystack 2.30 is here! Pass a plain string to any ChatGenerator
Maintained by deepset

Integration: OpenTelemetry

Trace and monitor your Haystack pipelines with OpenTelemetry.

Authors
deepset
deepset-ai
GitHub Repo PyPI Package

Table of Contents

Overview

This integration lets you use OpenTelemetry to trace and monitor your Haystack pipelines and agents. It builds on the OpenTelemetry SDK and provides an OpenTelemetryConnector component that, once added to your pipeline, sends Haystack traces to any OpenTelemetry-compatible backend.

Installation

pip install opentelemetry-haystack

Usage

Configure an OpenTelemetry TracerProvider with an exporter, then add the OpenTelemetryConnector to your pipeline without connecting it to any other component. It enables OpenTelemetry tracing for all pipeline operations.

You also need to set the HAYSTACK_CONTENT_TRACING_ENABLED environment variable to true to trace the content (inputs and outputs) of the pipeline components.

import os

os.environ["HAYSTACK_CONTENT_TRACING_ENABLED"] = "true"

from opentelemetry import trace
from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
from opentelemetry.sdk.resources import Resource
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor
from opentelemetry.semconv.resource import ResourceAttributes

# Configure the OpenTelemetry SDK. A service name is required for most backends.
resource = Resource(attributes={ResourceAttributes.SERVICE_NAME: "haystack"})
tracer_provider = TracerProvider(resource=resource)
tracer_provider.add_span_processor(BatchSpanProcessor(OTLPSpanExporter(endpoint="http://localhost:4318/v1/traces")))
trace.set_tracer_provider(tracer_provider)

from haystack import Pipeline
from haystack.components.builders import ChatPromptBuilder
from haystack.components.generators.chat import OpenAIChatGenerator
from haystack.dataclasses import ChatMessage

from haystack_integrations.components.connectors.opentelemetry import OpenTelemetryConnector

pipe = Pipeline()
pipe.add_component("tracer", OpenTelemetryConnector("Chat example"))
pipe.add_component("prompt_builder", ChatPromptBuilder())
pipe.add_component("llm", OpenAIChatGenerator(model="gpt-4o-mini"))

pipe.connect("prompt_builder.prompt", "llm.messages")

messages = [
    ChatMessage.from_system("Always respond in German even if some input data is in other languages."),
    ChatMessage.from_user("Tell me about {{location}}"),
]

response = pipe.run(
    data={"prompt_builder": {"template_variables": {"location": "Berlin"}, "template": messages}}
)
print(response["llm"]["replies"][0])

Your traces will then be available in the OpenTelemetry backend you configured (for example, Jaeger, Grafana Tempo, or any OTLP-compatible collector).

To capture even more detailed traces of the libraries used in your pipeline, you can combine the connector with the available OpenTelemetry instrumentations, such as opentelemetry-instrumentation-openai-v2 for OpenAI requests.

If you prefer not to use the connector, you can also enable the tracer manually:

from opentelemetry import trace
from haystack import tracing
from haystack_integrations.tracing.opentelemetry import OpenTelemetryTracer

tracing.enable_tracing(OpenTelemetryTracer(trace.get_tracer("my_application")))

License

opentelemetry-haystack is distributed under the terms of the Apache-2.0 license.