---
id: opentelemetry-instrumentation-sagemaker
version: "0.62.3"
license: Apache-2.0
license_treatment: permissive
maintenance: active
---

# opentelemetry-instrumentation-sagemaker — OpenTelemetry SageMaker instrumentation

License: permissive · Maintenance: active · Downloads: 6.8M/mo

## What it is and what it does

This package is an OpenTelemetry instrumentation plugin that automatically intercepts and traces calls to Amazon SageMaker endpoints made through Boto3. When activated, it wraps SageMaker API calls and records them as spans, capturing request bodies and response data by default to give visibility into model invocations and help debug LLM application behavior.

The instrumentation is designed for developers building observability into applications that invoke SageMaker models. It integrates with the OpenTelemetry ecosystem, requiring opentelemetry-api, opentelemetry-instrumentation, and semantic convention packages. By default it logs endpoint request and response content to span attributes; this can be disabled via the TRACELOOP_TRACE_CONTENT environment variable if privacy or trace size is a concern.

Use it for:

- Debug LLM application behavior by tracing what data is sent to and returned from SageMaker endpoints.
- Monitor SageMaker endpoint performance and latency in production observability systems.
- Evaluate output quality by capturing model responses in traces for later analysis.
- Reduce trace size in high-volume scenarios by disabling request/response body logging.
- Integrate SageMaker calls into a broader OpenTelemetry tracing pipeline across your application.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Automatically captures traces of Amazon SageMaker model invocations made through Boto3, recording request and response data as OpenTelemetry spans for observability and debugging.

Yes, if you are invoking SageMaker endpoints via Boto3 and need observability. The package is actively maintained, has no known vulnerabilities, installs cleanly, and integrates directly into OpenTelemetry. The privacy control via TRACELOOP_TRACE_CONTENT is a practical feature for sensitive workloads. Install it as part of your observability setup.

## Install

pip install opentelemetry-instrumentation-sagemaker
uv add opentelemetry-instrumentation-sagemaker
poetry add opentelemetry-instrumentation-sagemaker

## Installing opentelemetry-instrumentation-sagemaker

Before you install: Low friction install with a pure-Python wheel. Actively maintained as of 2026-08-10 with recent releases, and depends only on core OpenTelemetry packages.

License in practice: Apache-2.0 permissive license allows commercial and private use with minimal restrictions.

Quickstart:

pip install opentelemetry-instrumentation-sagemaker

from opentelemetry.instrumentation.sagemaker import SageMakerInstrumentor

SageMakerInstrumentor().instrument()

Requires Python 3.10 or later; requires an active OpenTelemetry trace provider to be configured separately.

Verify before relying:

- Whether the instrumentation works with all SageMaker endpoint types or only specific model types.
- Performance overhead of tracing on high-volume endpoint invocations.
- Compatibility with non-Boto3 SageMaker clients or SDKs.

## Package facts

- License: Apache-2.0 (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 6.8M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

sagemaker tracing, opentelemetry sagemaker, boto3 instrumentation, llm model tracing, sagemaker observability, endpoint request tracing, ai model monitoring, observability, aws-sagemaker, llm-tracing

[View on SkillFed](https://skillfed.io/packages/opentelemetry-instrumentation-sagemaker) · [View on PyPI](https://pypi.org/project/opentelemetry-instrumentation-sagemaker/)