Microsoft Fabric Activator generating repeated false alerts despite stable Power BI metric

Hi egstst,

Thanks for reaching out to the Microsoft Fabric Community forum.

Thanks for providing the detailed troubleshooting steps.

Activator queries the semantic model directly using the visual's underlying query and periodically evaluates the rule. As an additional troubleshooting step, you may want to review the configured query frequency for the Power BI data source, as Activator periodically evaluates rules based on the configured query interval.

For more details, please refer to the following offical documentation:

Activator ingestion from Power BI - Microsoft Fabric | Microsoft Learn

Query frequency for query data sources - Microsoft Fabric | Microsoft Learn

I hope this helps. Please feel free to reach out if you have any further questions.
Thank you.

Thank you for your response v-abhinavmu .

I understand that Activator periodically evaluates the semantic model based on the configured query frequency. I have tested both lower and higher query frequencies, but the issue persists—the only thing that changes is how frequently the false alerts are triggered. My concern is therefore not the evaluation frequency itself.

The semantic model is an Import model that refreshes three times each morning. After the final refresh, there are no further refreshes for the remainder of the day, and the report values remain stable.

Nevertheless, the Activator PowerBI Metric live feed continues to alternate between TRUE and FALSE for the same monitored objects, causing repeated "Changes to TRUE" alerts (with the alert frequency corresponding to the configured query frequency).

Since the underlying semantic model and report values remain unchanged after the final refresh, could you help explain why Activator is observing these alternating values?

• Is there a known issue with PowerBI Metric ingestion or evaluation for Import semantic models?
• Is there any diagnostic logging or a way to determine why Activator evaluates the metric as FALSE when the report continues to evaluate it as TRUE?

Many thanks,
Esther

Hi egstst,

Thank you for the update and for confirming that you've already tested different query frequencies.

Since the semantic model refreshes complete successfully, the report values remain stable, and the issue persists regardless of the configured query frequency, I'd recommend opening a Microsoft Support ticket for further investigation.

When creating the support request, it would be helpful to include:

• The Activator item and rule configuration.

• Semantic model refresh history.

• The configured query frequency.

• Screenshots showing the stable Power BI report alongside the Activator live feed alternating between TRUE/FALSE.

• Approximate timestamps of the repeated alerts.

This information will help the product team investigate the behavior further.

You can create a support ticket by following the steps in the Microsoft Learn article:

Create a Fabric and Power BI Support Ticket - Power BI | Microsoft Learn

Thank you.

Hi egstst,

Could you please confirm if the issue has been resolved after raising a support case? If a solution has been found, it would be greatly appreciated if you could share your insights with the community. This would be helpful for other members who may encounter similar issues.

Thank you for your understanding and assistance.

Yes refresh the page

v-abhinavmu I didn't end up opening a support ticket because the issue resolved itself around 29./30.06.2026. The setup has been stable since. I didn't change anything to resolve it...

If the issue starts again I will raise a ticket.

KenD very interesting, it seems that issue stopped for you around the same time...

Hi,

Based on your description, this sounds more like an issue with the Activator Power BI Metric event stream than with your semantic model or DAX measures.

A few observations:

• You've already confirmed the semantic model refreshes only once and the report values remain stable.

• Recreating the alerts and even changing the measures from 1/0 to TRUE/FALSE didn't change the behavior.

• The fact that the Activator live feed alternates between TRUE and FALSE while the report remains unchanged suggests Activator may be re-evaluating or replaying metric events rather than detecting actual changes.

A few things to check:

• Verify whether multiple report/dashboard tiles or duplicate Activator objects are monitoring the same metric.

• Review the Activator event history to see whether the repeated notifications correspond to new events or repeated evaluations of the same state.

• Test with a brand-new workspace, semantic model, and a simple Boolean measure to determine whether the issue is environment-specific or a broader service behavior.

If the issue started suddenly after months of working correctly and affects multiple alerts, it may indicate a recent service regression rather than a configuration problem.

I'd recommend opening a Microsoft Support ticket with:

• Workspace ID

• Activator item ID

• Semantic model ID

• Approximate timestamps of the repeated alerts

• Screenshots of the Activator live feed showing the alternating values

That information will allow the product team to inspect the backend event stream and determine whether duplicate or stale events are being generated.

If Microsoft confirms this as a known issue or provides a workaround, please consider sharing the resolution here. It would be very helpful for others experiencing the same behavior.

If you found this reply helpful, please consider giving it a Kudos. If it helps resolve your issue, marking it as the Accepted Solution will help other community members facing the same problem.

I had this same issue using a similar setup, and I tested by increasing the activator interval, and it seems to happen more often as it increases the event stream query within activator from 2 hours to 5 minutes. I had also tested by reducing the semantic refresh interval to almost one refresh, and the alert was still sending out false positives regularly. There was exactly one value different throughout the week, and it just exploded with alerts notification. The issue seems to have kind of stopped after 06/29 for me. I might test this again soon since I had to shut it down from the production pipeline and only run on Dev Semantic, with email being sent out specifically to me only. It would be great if someone could pinpoint why it starts and stops.

For me, the issue started throughout June (03-29). The screenshot below is one exampleDayWeek