Skip to main content
Haystack pipelines are instrumented by the OpenInference Haystack instrumentor, which captures pipeline runs, generator (model) calls, and retrieval steps as OTLP spans. Point the exporter at AgentMark and the traces arrive normalized. Haystack is a Python framework.

​
Setup

1

Install the instrumentor and the OTLP exporter

pip install openinference-instrumentation-haystack \
  opentelemetry-sdk opentelemetry-exporter-otlp-proto-http
2

Point the exporter at AgentMark and instrument Haystack

Use your AgentMark API key and app id from project settings.
from openinference.instrumentation.haystack import HaystackInstrumentor
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor
from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter

provider = TracerProvider()
provider.add_span_processor(
    BatchSpanProcessor(
        OTLPSpanExporter(
            endpoint="https://api.agentmark.co/v1/traces",
            headers={
                "Authorization": "<YOUR_API_KEY>",  # raw key, no "Bearer" prefix
                "X-Agentmark-App-Id": "<YOUR_APP_ID>",
            },
        )
    )
)

HaystackInstrumentor().instrument(tracer_provider=provider)
3

Run your pipeline

Run your Haystack pipeline as usual. Each pipeline run, generator call, and retrieval step arrives in AgentMark as a span, grouped into a trace. See Traces and logs.

​
What AgentMark captures

Haystack spans use the OpenInference attribute conventions — model, token usage, input and output messages, retrieval documents, tool calls, and span kind are all mapped onto AgentMark’s normalized trace fields, and token counts feed cost tracking. See OpenInference for the full attribute mapping.

​
Next steps

OpenInference

How AgentMark reads OpenInference attributes

Traces and logs

Explore traces once they arrive

​
Have questions?

Reach out any time: