fix: create root span per prompt for OTEL trace isolation

[josephwangrb]

Issue for this PR

Closes #32920

Type of change

• Bug fix

What does this PR do?

When OTEL_EXPORTER_OTLP_ENDPOINT is set, all spans across all prompts in a session inherit the server's boot-time trace context, producing one giant trace per session instead of one trace per prompt. This also causes the <trace-without-root-span> issue reported in #32920 — there's no root span boundary per prompt, so backends like Jaeger can't group spans into meaningful traces.

Root cause: SessionPrompt.loop was defined as Effect.fn("SessionPrompt.loop"), which creates a child span of the ambient context — not a new root span. The Effect runtime's AsyncLocalStorageContextManager propagates the trace context from server boot, so every Effect.fn() call inherits the same trace ID.

Fix: Wrap SessionPrompt.loop in Effect.withSpan("SessionPrompt.loop", { root: true }) so each prompt starts a new trace. All child spans (SessionProcessor.process, LLM.run, ai.streamText, tool execution) then inherit the new trace ID.

// Before:
const loop = Effect.fn("SessionPrompt.loop")(function* (input) {
  return yield* state.ensureRunning(input.sessionID, lastAssistant(input.sessionID), runLoop(input.sessionID))
})

// After:
const loop = (input: LoopInput) =>
  state
    .ensureRunning(input.sessionID, lastAssistant(input.sessionID), runLoop(input.sessionID))
    .pipe(Effect.withSpan("SessionPrompt.loop", { root: true, attributes: { "session.id": input.sessionID } }))

The { root: true } option tells Effect's tracer to create a new root span (new trace ID) instead of parenting to the ambient context.

How did you verify your code works?

Built opencode with the fix and tested via opencode serve + SDK with 2 prompts (4 turns total) against a local OTel collector:

Before — 1 trace for entire session (1429 spans, 30.4s), all turns in one trace.

After — 1 trace per prompt:

• Prompt 1: trace 195697cf... (488 spans) with 2 SessionProcessor.process turns and ai.streamText.doStream carrying gen_ai.request.model=glm-5.2, token usage, and finish reason
• Prompt 2: trace 82b2b0cd... (450 spans) with 2 turns, same GenAI attributes

Each SessionPrompt.loop span has parent=(none) confirming it's a root span. Typecheck passes (bun turbo typecheck — 30/30 packages).

Checklist

• I have tested my changes locally
• I have not included unrelated changes in this PR

[github-actions]

Thanks for your contribution!

This PR doesn't have a linked issue. All PRs must reference an existing issue.

Please:

1. Open an issue describing the bug/feature (if one doesn't exist)
2. Add Fixes #<number> or Closes #<number> to this PR description

See CONTRIBUTING.md for details.

[github-actions]

The following comment was made by an LLM, it may be inaccurate:

Found a potentially related PR:

Related PR:

• feat(observability): add v2 genai tracing #35935 — feat(observability): add v2 genai tracing
  • Related because it addresses observability and GenAI tracing, which overlaps with the trace span management improvements in PR fix: create root span per prompt for OTEL trace isolation #36179. Depending on the scope of feat(observability): add v2 genai tracing #35935, it may have similar trace context handling logic that should be coordinated.

The other result (#5245 — "feat: integrate OpenTelemetry") appears to be an older PR and may not be directly relevant to the current span isolation fix.

[github-actions]

Thanks for updating your PR! It now meets our contributing guidelines. 👍