Bulk exporting trace data delays

There has been several hour of delay on bulk exports from langsmith. I looked at the status board and aside from instability a few days ago everything looks fine.

Usually they are available within 15 minutes of the hour’s completion. Any ideas?

I would also like to know if there is an SLO for this?

Hi @myusuf3

On the export delay

Bulk exports are typically processed asynchronously and can be delayed if:

  • There’s elevated load on the export workers

  • Large export volumes are queued ahead of yours

  • There was recent instability (even if current status is green, backlogs can linger)

If you’re seeing several hours of delay (vs the usual ~15 minutes after the hour), that’s longer than typical behavior and likely indicates a temporary backlog rather than normal variance.


Is there an SLO?

As of now, there isn’t a publicly documented SLO specifically for bulk export completion latency.

Exports are generally best-effort background jobs rather than real-time guarantees. If you depend on them for production workflows (e.g., compliance pipelines, data warehouse ingestion), it’s worth raising that with support to clarify expectations for your use case.


What I’d Suggest

  1. Confirm:

    • Export size (row count / time range)

    • Whether this affects all exports or just specific large ones

  2. Check:

    • Whether newer exports are also delayed

    • Or if only a specific batch is stuck

  3. If the delay persists:

    • Open a support ticket with:

      • Workspace/org ID

      • Export job IDs

      • Approximate submission times

That helps the team quickly inspect queue state.