feat(traceloop-sdk): add trace_content parameter to Traceloop.init()

[smz202000]

Summary

Adds a trace_content parameter to Traceloop.init() that allows users to control whether sensitive content (prompts and completions) is sent in traces.

Fixes #137

Changes

• Modified is_content_tracing_enabled() in traceloop/sdk/config/__init__.py to accept an optional trace_content parameter. When explicitly provided, it overrides the TRACELOOP_TRACE_CONTENT environment variable. When not provided (None), it falls back to the environment variable (backward compatible).
• Added trace_content: Optional[bool] = None parameter to Traceloop.init() in traceloop/sdk/__init__.py.
• Added docstring documentation for the new parameter.

Usage

# Disable content tracing programmatically (overrides env var)
Traceloop.init(app_name="my-app", trace_content=False)

# Enable content tracing explicitly (overrides env var)
Traceloop.init(app_name="my-app", trace_content=True)

# Default behavior: uses TRACELOOP_TRACE_CONTENT env var
Traceloop.init(app_name="my-app")

Backward Compatibility

This change is fully backward compatible. When trace_content is not passed (the default), the existing TRACELOOP_TRACE_CONTENT environment variable behavior is preserved.

Summary by CodeRabbit

• New Features
  • Added an optional setting to control whether prompt and completion content is included in traces.
  • Added configuration options for enabling or disabling metrics and logging.
  • Content tracing now supports explicit configuration while retaining environment-based defaults.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 074553ce-9efa-4b86-ae67-dd7296ee7862

📥 Commits

Reviewing files that changed from the base of the PR and between 93429cf and 0da9614.

📒 Files selected for processing (2)

• packages/traceloop-sdk/traceloop/sdk/__init__.py
• packages/traceloop-sdk/traceloop/sdk/config/__init__.py

---

📝 Walkthrough

Walkthrough

Traceloop.init() now accepts an optional trace_content override. Content tracing falls back to TRACELOOP_TRACE_CONTENT when unset, and new helpers read metrics and logging enablement from environment variables.

Changes

Trace content configuration

Layer / File(s)	Summary
Configuration resolution helpers packages/traceloop-sdk/traceloop/sdk/config/__init__.py	Content tracing supports an optional boolean override, while metrics and logging enablement are derived from their respective environment variables.
Traceloop initialization integration packages/traceloop-sdk/traceloop/sdk/__init__.py	Traceloop.init() exposes and documents trace_content, then passes it to content-tracing configuration.

Estimated code review effort: 2 (Simple) | ~10 minutes

Suggested reviewers: dvirski

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Out of Scope Changes check	⚠️ Warning	The new metrics and logging enablement helpers are unrelated to issue #137 and appear outside the requested scope.	Move the metrics/logging helpers to a separate PR or justify them in the issue scope before merging.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely describes the main change: adding a trace_content parameter to Traceloop.init().
Linked Issues check	✅ Passed	The PR implements the requested trace_content override for Traceloop.init() and preserves env-var fallback when unset.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

✨ 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.}

[GioiaZheng]

Independent privacy review against 0da96148: the SDK-level override does not currently reach the actual instrumentation gates.

I ran the same matrix against TracerWrapper.enable_content_tracing, OpenAI's should_send_prompts(), and the SDK decorator gate:

env	arg	SDK	OpenAI/decorator
unset	None	true	true
true	False	false	true
false	True	true	false
false	None	false	false
true	None	true	true

Minimal regression check:

monkeypatch.setenv("TRACELOOP_TRACE_CONTENT", "true")
Traceloop.init(
    exporter=InMemorySpanExporter(),
    disable_batch=True,
    trace_content=False,
)
assert TracerWrapper.enable_content_tracing is False
assert should_send_prompts() is False  # currently True

I also exercised an OpenAI chat span with an in-memory exporter and a mock HTTP transport (no provider call). With env=true, trace_content=False, the emitted openai.chat span still contained both the synthetic prompt and completion, including gen_ai.input.messages, although the SDK flag was false. The inverse conflict (env=false, trace_content=True) also leaves the instrumentation disabled.

The existing SDK privacy tests pass (29/29) because they disable content through the environment variable and therefore do not cover this split-brain case. I think this needs an end-to-end regression test at the instrumentation/span boundary before merge; changing only is_content_tracing_enabled() is not sufficient while provider instrumentations and decorators independently read TRACELOOP_TRACE_CONTENT.