Forum Discussion

V_S27's avatar
V_S27
Frequent Visitor
1 year ago

Automatic Aggregations wiped out with error

Hello,

 

I have automatic aggregations setup for my semantic model & they have been working perfectly fine. However, suddenly it has started showing the following error -

"No SystemManaged table found after PA training"

I have almost 3 GB of aggregations till yesterday and today its all wiped out. No changes to the semantic model were published & we have daily refreshes scheduled. Can someone help me understand why this happened & how to resolve it.

Thanks!

Need Help

5 Replies

  • v-hjannapu's avatar
    v-hjannapu
    Icon for Community Support rankCommunity Support
    1 year ago

    Hi V_S27,

    Thank you  for reaching out to the Microsoft fabric community forum.

    That error message "No SystemManaged table found after PA training" usually means that Power BI tried to run automatic aggregations but didn’t end up creating any this time. That’s likely why your 3GB of aggregation data disappeared. This can happen even if you didn’t publish any model changes things like shifts in how users are querying the data, updates to the source schema, or memory pressure during refresh can cause it.

    I would suggest opening the Aggregations tab in your semantic model settings, checking the last training result, and clicking “Retrain” to try running it again. It’s also worth looking at your refresh logs and capacity metrics to see if something unusual happened during the last run. If this keeps happening, setting up manual aggregations for your key queries might help keep things stable.

    Kindly refer to the below documentation links for better understanding:
    https://learn.microsoft.com/en-us/fabric/enterprise/powerbi/aggregations-auto-configure

    Hope this helps if you have any queries we are  happy to assist you further.
    Best Regards,
    Harshitha.

    • V_S27's avatar
      V_S27
      Frequent Visitor
      1 year ago

      Hi v-hjannapu ,

      Thanks for the quick response. We did check the capacity metrics and evrything seems ok. There was no throttling and the capcity usage was as usual. I tried re-training it but still gives me the same error. How can I forsee such situations, there is no way to predict if this will happen again. This is a huge inconvenience actually.

      Thanks!

  • Shahid12523's avatar
    Shahid12523
    Icon for Community Champion rankCommunity Champion
    1 year ago

    The error means system-managed aggregation tables got wiped (often from refresh failure, schema drift, or capacity reset).

    Fix:

    Retrain or toggle Automatic Aggregations in model settings.

    Check capacity logs for throttling.

    Verify source schema.

    If it keeps happening → raise a Microsoft support ticket.