Relari-OTel

The `relari-otel package provides a lightweight wrapper around OpenTelemetry primitives for tracing LLM applications and send the traces to a collector. This repository is met to be used with Agent-Contracts

Quickstart

First, install the package, specifying the instrumentations you want to use.

For example, to install the package with OpenAI and Langchain/Langgraph instrumentations:

pip install 'relari-otel[langchain,openai]'

Initialize the telemetry in your code

Relari.init(project_name="Your Project Name")

Certification

If you intend to use the certification feature (see agent-contracts docs) you need to add the vertification extra:

pip install 'relari-otel[langchain,openai,certification]'

Instrumentation

All your framework and models are instrumented automatically. However OpenTelemetry does not know how your code is structured. To wrap all the spans in a trace you can use

with Relari.start_new_sample(scenario_id="scenario_A"):
    await my_agent_code(specs["scenario_A"]["data"])

or alternative to run an evaluation over a specification

async def runnable(data: Any):
    agent_inputs = prepare_my_agent_input(data)
    return my_agent_code(agent_inputs)

await Relari.eval_runner(specs=specs, runnable=runnable)

Acknowledgements

Autoinstrumentations are provided by Arize