---
title: python-agent-110200
source: https://docs.newrelic.com/docs/release-notes/agent-release-notes/python-release-notes/python-agent-110200
---

## Notes

This release of the Python agent adds support for the following:

-   [Strands](https://pypi.org/project/strands-agents-tools/)
-   [AWS Bedrock Converse Streaming]((https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/bedrock-runtime/client/converse.html))
-   [Redis](https://pypi.org/project/redis/) methods `hybrid_search` and `msetex`
-   [Langchain](https://pypi.org/project/langchain/) role assignments for LLM events
-   LLM event timestamp captured at time of LLM request
-   `loop_factory` parameter in `asyncio.Runner`

This release also addresses a bug in ASGI instrumentation where headers are not reported by the web server.

Install the agent using `easy_install/pip/distribute` via the [Python Package Index](https://pypi.python.org/pypi/newrelic), using `conda` via the [Conda-Forge Package Index](https://anaconda.org/conda-forge/newrelic), or download it directly from the [New Relic download site](https://download.newrelic.com/python_agent/release).

## Features

-   Add support for [Strands](https://pypi.org/project/strands-agents-tools/)

    -   Add instrumentation for Strands tools and agents in single and multi-agent (swarms, graphs, etc.) setups

-   Add support for [AWS Bedrock Converse Streaming](https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/bedrock-runtime/client/converse.html)

    -   Add monitoring support for streaming Converse API. This includes support for Converse chat completion calls made with boto3 and aioboto3 clients using `converse_stream()`.

-   Add support for [Redis](https://pypi.org/project/redis/) methods

    -   Add support for the following methods:
    -   `hybrid_search`
    -   `msetex`

-   Add support for `loop_factory` in `asyncio.Runner`

    -   In Python 3.11, the `asyncio.Runner` class was added as well as the `loop_factory` parameter. The `loop_factory` parameter provides a replacement for loop policies.  Now, the loop that is created is instrumented, regardless of whether it is through a global event loop setting or a custom event loop.

-   Capture [Langchain](https://pypi.org/project/langchain/) role assignments in LLM events

    -   Langchain `LlmChatCompletionMessage` events now contains the assigned role from [Langchain](https://pypi.org/project/langchain/).

-   Capture LLM event timestamp at time of LLM request

    -   Previously, the LLM event's timestamp was captured at the time of event creation. Now, it is more accurately captured at the time of the LLM request.

## Bug fixes

-   ASGI header consumption

    -   When an `ASGIWebTransaction` consumes the headers from a `http.response.start` event, it was doing so without regard to the allowable input types. When the headers are a generator, the generator was consumed by the call to `process_response()` leaving no headers remaining for the web server to send.  This has now been fixed.

## Support statement

We recommend updating to the latest agent version as soon as it's available. If you can't upgrade to the latest version, update your agents to a version no more than 90 days old. Read [more](https://docs.newrelic.com/docs/new-relic-solutions/new-relic-one/install-configure/update-new-relic-agent/) about keeping agents up to date.

See the New Relic Python agent [EOL policy](https://docs.newrelic.com/docs/apm/agents/python-agent/getting-started/python-agent-eol-policy/) for information about agent releases and support dates.