Destinations overview

Send AI observability data from OpenLIT to your existing observability stack using standard OTLP

The OpenLIT SDK sends traces, metrics, and logs using the OpenTelemetry Protocol (OTLP) over HTTP. Any backend that accepts OTLP data — whether a managed SaaS service or a self-hosted stack — can receive data from OpenLIT without any vendor-specific plugins or agents.

How it works

Configure the destination endpoint and authentication headers in openlit.init() or via environment variables:

Python

import openlit

openlit.init(
    otlp_endpoint="https://your-otlp-endpoint",
    otlp_headers="Authorization=Bearer your-token"
)

TypeScript

import Openlit from 'openlit';

Openlit.init({
    otlpEndpoint: 'https://your-otlp-endpoint',
    otlpHeaders: 'Authorization=Bearer your-token'
});

Environment variables

export OTEL_EXPORTER_OTLP_ENDPOINT="https://your-otlp-endpoint"
export OTEL_EXPORTER_OTLP_HEADERS="Authorization=Bearer your-token"

Then initialize with no arguments:

import openlit
openlit.init()

If you do not set otlp_endpoint, the SDK outputs traces to the console. This is useful during development but should not be used in production.

Supported destinations

OpenLIT

Send to a self-hosted OpenLIT platform instance. Default OTLP endpoint: http://127.0.0.1:4318.

Grafana Cloud

Send to the Grafana Cloud OTLP gateway using Basic Auth with your Instance ID and API token.

Datadog

Send to a local Datadog Agent that forwards traces and metrics to the Datadog backend.

OpenTelemetry Collector

Send to any OpenTelemetry Collector instance for fan-out to multiple backends simultaneously.

Configuration reference

Parameter	Environment Variable	Description
otlp_endpoint	OTEL_EXPORTER_OTLP_ENDPOINT	OTLP HTTP endpoint URL
otlp_headers	OTEL_EXPORTER_OTLP_HEADERS	Authentication headers as key=value pairs, comma-separated
service_name	OTEL_SERVICE_NAME	Service name shown in your backend
environment	OTEL_DEPLOYMENT_ENVIRONMENT	Deployment environment (e.g., production)

Environment variables take precedence over parameters passed to openlit.init(). This lets you manage destination configuration outside your application code, which is recommended for production deployments.

---

Integrations

60+ LLM providers, AI frameworks, and vector databases

Configuration

Full reference for all openlit.init() parameters

GPU integrations

Monitor NVIDIA and AMD GPU metrics