traceloop
🚀 Feature: allow disabling prompt sending as an argument to Traceloop.init()
Which component is this feature for?
Traceloop SDK
🔖 Feature description
Add a setting called trace_content (which will default to true) and will control whether all the instrumentations are sending sensitive content (like prompts and completions). This should override the env var behavior (similar to other flags we have).
🎤 Why is this feature needed ?
More control on this sensitive feature
✌️ How do you aim to achieve this?
🔄️ Additional Information
No response
👀 Have you spent some time to check if this feature request has been raised before?
- [X] I checked and didn't find similar issue
Are you willing to submit PR?
None
@nirga @randolfuy01 Hi there! Is it still WIP? If not, I would love to start looking around this issue:)
Not sure what's @randolfuy01's status, but there are other issues that you can also look at that are unassigned @sanspareilsmyn!
@nirga Hi! Is this issue still open? If so, I would be happy to look at it :)
@nirga To clarify in greater detail,
- The setting is the flag in the sdk's config init.py 'is_content_tracing_enabled
https://github.com/traceloop/openllmetry/blob/9ad18b7067f3036a13268b50366ff4207e32ffa2/packages/traceloop-sdk/traceloop/sdk/config/init.py#L8-L9
- Then the setting is called in the global sdk init.py
https://github.com/traceloop/openllmetry/blob/9ad18b7067f3036a13268b50366ff4207e32ffa2/packages/traceloop-sdk/traceloop/sdk/init.py#L79
- And I need to write the functionality that for each instrument,
https://github.com/traceloop/openllmetry/blob/9ad18b7067f3036a13268b50366ff4207e32ffa2/packages/traceloop-sdk/traceloop/sdk/init.py#L55
the sensitive content being shared for each individual instrument is controlled.