opentelemetry-instrumentation-vertexai

What it is and what it does

This package is an OpenTelemetry instrumentation plugin that automatically intercepts calls to the official VertexAI library and records them as distributed traces. When you call VertexAI APIs for prompts, completions, or embeddings, the instrumentation captures the request and response data and exports it as OpenTelemetry spans, which can then be sent to any compatible tracing backend.

By default, the instrumentation logs the full content of prompts, completions, and embeddings into span attributes, giving you visibility into what your LLM application is sending and receiving. This is useful for debugging and evaluating output quality, but you can disable it by setting the TRACELOOP_TRACE_CONTENT environment variable to false if the data contains sensitive user information or if you want to reduce trace size.

Use it for:

• Debug LLM application behavior by inspecting prompts and completions in a distributed trace viewer.
• Monitor VertexAI API latency and error rates across your application using OpenTelemetry backends.
• Audit and log all AI API calls for compliance or quality assurance without modifying application code.
• Correlate VertexAI calls with other service traces to understand end-to-end request flow.
• Reduce trace verbosity by disabling content logging while keeping call metadata and performance metrics.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Automatically captures traces of VertexAI API calls—prompts, completions, and embeddings—and exports them as OpenTelemetry spans for observability and debugging.

Yes, if you use VertexAI and already have an OpenTelemetry setup or plan to adopt one. The package is actively maintained, has no known vulnerabilities, and integrates seamlessly with the OpenTelemetry ecosystem. Install friction is low. The privacy control via environment variable is a practical feature for production use. No significant blockers.

Install

opentelemetry-instrumentation-vertexai on PyPI

pip install opentelemetry-instrumentation-vertexai

uv add opentelemetry-instrumentation-vertexai

poetry add opentelemetry-instrumentation-vertexai

Installing opentelemetry-instrumentation-vertexai

Before you install

Low friction: pure Python wheel with four lightweight OpenTelemetry dependencies. Actively maintained as of 2026-08-10 with recent release history.

License in practice

Apache-2.0 permissive license allows commercial and private use with minimal restrictions.

Quickstart

pip install opentelemetry-instrumentation-vertexai

from opentelemetry.instrumentation.vertexai import VertexAIInstrumentor

VertexAIInstrumentor().instrument()

Requires Python 3.10 or later; VertexAI library must be installed separately to instrument.

Verify before relying

• Whether trace export backends (e.g., Jaeger, Datadog, GCP Cloud Trace) are automatically configured or require separate setup.
• Performance overhead when tracing high-volume VertexAI requests.
• Compatibility with specific VertexAI library versions beyond the stated Python range.

Package facts

Evidence: opentelemetry_instrumentation_vertexai-0.62.3-py3-none-any.whl

Tags

More Monitoring packages