opentelemetry-instrumentation-botocore

What it is and what it does

This package hooks into botocore and aiobotocore to emit OpenTelemetry spans for every AWS API call your application makes. It integrates with the OpenTelemetry ecosystem—using opentelemetry-api, opentelemetry-instrumentation, and opentelemetry-semantic-conventions—to standardize how AWS request traces are recorded and exported to observability backends.

It includes specialized extensions for several AWS services: DynamoDB, Lambda, SNS, and SQS get automatic semantic enrichment, while Bedrock Runtime gets GenAI semantic conventions for Converse, ConverseStream, InvokeModel, and InvokeModelWithResponseStream calls. The Bedrock extension supports tracing for Amazon Titan, Amazon Nova, and Anthropic Claude models, with tool-call tracing for Nova and Claude 3+. Once installed and activated, instrumentation happens transparently via wrapt.

Use it for:

• Trace all DynamoDB operations in a microservice to identify slow queries or hot partitions in your observability backend.
• Monitor Lambda invocations triggered by your application and correlate them with parent traces across service boundaries.
• Capture SNS/SQS message publish and receive operations to debug asynchronous workflows and message delivery latency.
• Instrument Bedrock API calls to track GenAI model invocations, token usage, and tool calls in a centralized trace store.
• Debug AWS SDK errors and timeouts by examining detailed span attributes and events without modifying application code.

Worth the install?

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

Adds distributed tracing to AWS SDK calls made through botocore and aiobotocore, capturing request/response telemetry for services like DynamoDB, Lambda, SNS, SQS, and Bedrock Runtime.

Yes, if you are already using OpenTelemetry in your application and need to trace AWS SDK calls. The package is actively maintained, has no known vulnerabilities, and integrates cleanly into the OpenTelemetry ecosystem. It is still in beta (0.65b0), so expect occasional API changes; production use is reasonable but monitor release notes. Skip it if you are not using OpenTelemetry or do not need AWS call tracing.

Install

opentelemetry-instrumentation-botocore on PyPI

pip install opentelemetry-instrumentation-botocore

uv add opentelemetry-instrumentation-botocore

poetry add opentelemetry-instrumentation-botocore

Installing opentelemetry-instrumentation-botocore

Before you install

Low friction installation with a pure-Python wheel. Actively maintained with a recent release (29 days old) and steady commit activity. Depends on five core OpenTelemetry packages, all standard in the observability ecosystem.

License in practice

Apache-2.0 permissive license allows use in commercial and proprietary projects with minimal restrictions; attribution required but no copyleft obligations.

Quickstart

pip install opentelemetry-instrumentation-botocore

from opentelemetry.instrumentation.botocore import BotocoreInstrumentor
BotocoreInstrumentor().instrument()

Requires Python 3.10 or later; assumes an active OpenTelemetry tracer provider is configured in your application.

Verify before relying

• Whether Bedrock tracing metrics/events work for all model families or only the explicitly listed subset (Titan, Nova, Claude).
• Whether tool-call tracing for InvokeModel APIs extends to models beyond Nova and Claude 3+.
• Performance overhead of instrumentation on high-volume AWS SDK call workloads.
• Compatibility with aiobotocore async workflows and any async-specific tracing behavior.

Package facts

Evidence: opentelemetry_instrumentation_botocore-0.65b0-py3-none-any.whl

Tags

More Monitoring packages