Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

[BOT ISSUE] Google GenAI (Gemini) streaming calls are not instrumented

Open
#61 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
48/100
Issue type
Feature
Clarity
Mostly clear
Activity status
Quiet
Tech stack
java
Domain
observability

Research direction

Start with BraintrustApiClient.java lines 186–312 and inspect the upstream ApiClient streaming methods used by generateContentStream. Compare the SSE handling and chunk accumulation in the OpenAI and Anthropic TracingHttpClient implementations, then review BraintrustGenAITest.java. Done means streaming Gemini calls produce spans, output capture, and token usage metrics.

Written by the indexing model from the issue text.

Description

Summary

The Google GenAI instrumentation wraps ApiClient.request() and ApiClient.asyncRequest() for synchronous and asynchronous non-streaming calls, but does not override any streaming methods. Calls to client.models.generateContentStream() and similar streaming APIs bypass the instrumentation entirely, producing no spans, no output capture, and no token usage metrics.

What is missing

BraintrustApiClient.java overrides 4 methods:

  • request(String, String, String, Optional) (line 186)
  • request(String, String, byte[], Optional) (line 210)
  • asyncRequest(String, String, String, Optional) (line 239)
  • asyncRequest(String, String, byte[], Optional) (line 277)

The upstream ApiClient class in the Google GenAI SDK also has streaming request methods (used internally by generateContentStream) that are not overridden. This means streaming Gemini calls — a common pattern for chat applications — are completely invisible to Braintrust tracing.

For comparison, both the OpenAI and Anthropic instrumentations in this repo handle streaming via SSE detection and chunk accumulation in their respective TracingHttpClient implementations.

Upstream sources

Braintrust docs status

  • not_found — braintrust.dev/docs does not reference Google GenAI streaming instrumentation for Java. The Java SDK docs page redirects to the GitHub repo.

Local files inspected

  • braintrust-sdk/instrumentation/genai_1_18_0/src/main/java/com/google/genai/BraintrustApiClient.java — lines 186–312 (only non-streaming request/asyncRequest overridden; no streaming method override)
  • braintrust-sdk/instrumentation/genai_1_18_0/src/main/java/com/google/genai/BraintrustInstrumentation.java — wraps ApiClient in Client and service fields, no streaming awareness
  • braintrust-sdk/instrumentation/genai_1_18_0/src/main/java/dev/braintrust/instrumentation/genai/v1_18_0/BraintrustGenAI.java — thin wrapper, delegates to BraintrustInstrumentation
  • braintrust-sdk/instrumentation/genai_1_18_0/src/test/java/dev/braintrust/instrumentation/genai/v1_18_0/BraintrustGenAITest.java — only tests sync and async non-streaming; no streaming tests exist
Dominant language
Java
Stars
21
Forks
5
Avg merge
2d 7h
Merged PRs (30d)
8

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from braintrustdata/braintrust-sdk-java

All issues in braintrustdata/braintrust-sdk-java

Similar issues

More Java issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.