Tracking (logging) manual updates in dataset items

[fleschadam]

I am using self-hosted version. I need a way to track (audit log) manual changes made to any dataset items. And i'd need specifically the timestamp when the last saving happened on the item which should become a metadata attaced to the item. Thanks for your help!

[inkeep-ai-assistant[bot]]

---

Based on the available documentation, I cannot find specific information about built-in audit logging capabilities for manual dataset item changes in Langfuse's self-hosted version.

However, the documentation shows that dataset items support metadata fields(1), which could be used to implement your own tracking mechanism. You can update dataset items with metadata using the SDK:

langfuse.create_dataset_item(
 id="<item_id>",
 # example: update status to "ARCHIVED"
 status="ARCHIVED"
)

(2)

Since dataset items support arbitrary JSON metadata(1), you could potentially implement a custom solution by:

1. Adding timestamp metadata when updating items programmatically
2. Using the metadata field to store change history
3. Implementing application-level logging for manual UI changes

For comprehensive audit logging of manual UI changes, you might need to check if your self-hosted deployment has access to application logs or consider implementing custom tracking through the API endpoints.

The most relevant resources for your use case would be:

• Datasets documentation
• API reference for dataset operations

Would you like a general approach for implementing custom audit logging for dataset changes?

📚 Sources:

1. Datasets v2 - Langfuse
2. Datasets

---

Have another question? Just tag @inkeep.

[jannikmaierhoefer]

Hi @fleschadam, this is part of our self-hosted enterprise license: https://langfuse.com/docs/administration/audit-logs