fix: flush tracer on WS responses final chunk

[thiscantbeserious]

Summary

For native WebSocket streaming, the postHookRunner closure returned by RunStreamPreHooks called drainAndAttachPluginLogs on the final chunk but never called tracer.CompleteAndFlushTrace. The logging plugin accumulates completed entries in an in-memory pendingLogsToInject map keyed by trace ID. That map is only drained to the store when CompleteAndFlushTrace runs. Every native-WS response therefore left its log entry stuck in memory and no row was written to the logs table.

Root cause: a single missing CompleteAndFlushTrace call in the happy-path branch of RunStreamPreHooks. All early-exit paths in the same function and the sibling RunRealtimeTurnPreHooks already call it correctly. The HTTP streaming path also flushes correctly.

Changes

• core/bifrost.go: add tracer.CompleteAndFlushTrace(traceID) inside the postHookRunner closure when IsFinalChunk(ctx) is true, after drainAndAttachPluginLogs. Mirrors the pattern used on every other exit path in the same function.
• core/bifrost_test.go: add TestRunStreamPreHooks_PostHookRunner_FlushesTracerOnFinalChunk and TestRunStreamPreHooks_PostHookRunner_NoFlushOnNonFinalChunk using a recording tracer to assert the call happens exactly once on the final chunk and not on mid-stream chunks.

Type of change

• Bug fix

Affected areas

• Core (Go)

How to test

go build ./core/... ./transports/...
go test ./core/... -run TestRunStreamPreHooks
go vet ./core/...

Expected: both new tests pass, build and vet are clean.

End-to-end: start Bifrost, route a WebSocket Responses request through a mock upstream that sends a terminal response.completed frame, then check the logs table. Before this fix the row was never written. After this fix a row appears.

Screenshots/Recordings

N/A

Breaking changes

• No

Related issues

Closes #2999

This bug was discovered while working on PR #2775 (OpenAI OAuth passthrough). The PR branch contains the fix that is being extracted here.

Security considerations

None. This change only adds a missing trace flush call.

Checklist

• I read docs/contributing/README.md and followed the guidelines
• I added/updated tests where appropriate
• I updated documentation where needed
• I verified builds succeed (Go and UI)
• I verified the CI pipeline passes locally if applicable

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: ace8ec19-7301-4f54-be50-a2476c29008b

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[thiscantbeserious]

Superseded by #3018 (merged 2026-04-24), which bundles the fix for this issue and several other native WS reliability bugs. Closing this draft as redundant.