botocore: add basic tracing for bedrock ConverseStream

[xrmx]

Description

This introduces basic tracing for the ConverseStream service. I've chosen to delay the closing of the span after the response has been consumed instead of buffering all the response chunks since the buffer it may become big.

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide instructions so we can reproduce. Please also list any relevant details for your test configuration.

• tox -e py310-test-instrumentation-botocore

Does This PR Require a Core Repo Change?

• Yes. - Link to PR:
• No.

Checklist:

See contributing.md for styleguide, changelog guidelines, and more.

• Followed the style guidelines of this project
• Changelogs have been updated
• Unit tests have been added
• Documentation has been updated

[codefromthecrypt]

Looks good to me. Do you mind checking if this langchain code produces spans as intended?

from langchain_aws import ChatBedrock


def main():
    llm = ChatBedrock(
        beta_use_converse_api=True,
        model_id='amazon.titan-text-lite-v1',
        streaming=True,
    )
    response_stream = llm.stream("Write a short poem on OpenTelemetry.")
    for chunk in response_stream:
        for item in chunk.content:
            if 'text' in item:
                print(item['text'], end="")

if __name__ == "__main__":
    main()

[xrmx]

from langchain_aws import ChatBedrock


def main():
    llm = ChatBedrock(
        beta_use_converse_api=True,
        model_id='amazon.titan-text-lite-v1',
        streaming=True,
    )
    response_stream = llm.stream("Write a short poem on OpenTelemetry.")
    for chunk in response_stream:
        for item in chunk.content:
            if 'text' in item:
                print(item['text'], end="")

if __name__ == "__main__":
    main()

It does:

{
    "name": "chat amazon.titan-text-lite-v1",
    "context": {
        "trace_id": "0xa584c79bfbd24ee4a69a470f62f5a785",
        "span_id": "0x330a7997ae2fd3c8",
        "trace_state": "[]"
    },
    "kind": "SpanKind.CLIENT",
    "parent_id": null,
    "start_time": "2025-01-24T08:43:40.741219Z",
    "end_time": "2025-01-24T08:43:44.649476Z",
    "status": {
        "status_code": "UNSET"
    },
    "attributes": {
        "rpc.system": "aws-api",
        "rpc.service": "Bedrock Runtime",
        "rpc.method": "ConverseStream",
        "aws.region": "us-east-1",
        "gen_ai.system": "aws.bedrock",
        "gen_ai.request.model": "amazon.titan-text-lite-v1",
        "gen_ai.operation.name": "chat",
        "aws.request_id": "c41388e1-9406-4376-8d43-8df022879225",
        "retry_attempts": 0,
        "http.status_code": 200,
        "gen_ai.usage.input_tokens": 12,
        "gen_ai.usage.output_tokens": 128,
        "gen_ai.response.finish_reasons": [
            "max_tokens"
        ]
    },
    "events": [],
    "links": [],
    "resource": {
        "attributes": {
            "telemetry.sdk.language": "python",
            "telemetry.sdk.name": "opentelemetry",
            "telemetry.sdk.version": "1.30.0.dev0",
            "telemetry.auto.version": "0.51b0.dev",
            "service.name": "unknown_service"
        },
        "schema_url": ""
    }
}

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: xrmx / name: Riccardo Magliocchetti (de33ce2, bd8a50f, aeab196, c6a7897, 2a3436a, 35de237, 0a896f6)

[lmolkova]

what if stream terminates with an error? what if user closes it before receiving response fully? We should try to cover as much as we can and I believe OpenAI instrumentation does it

[xrmx]

I'll cover more cases in followup PRs, thanks for reminding about this.

[xrmx]

Added in the checklist to not forgetting #3210