anomalies are a work in progress.

docs/use-cases/observability/clickstack/demo-days/2026/clickstack-new-features-05-06-2026.md

@@ -50,9 +50,9 @@
 
 <iframe width="768" height="432" src="https://www.youtube.com/embed/aAgN3ZlQz8M" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" allowfullscreen></iframe>
 
-ClickStack now supports anomaly-based alerting using Z-score detection. Instead of a fixed threshold, you set how many standard deviations from the expected mean should trigger an alert. The seasonality window (hourly or daily) controls how the baseline is computed, and the alert editor highlights the windows that would have fired so you can tune sensitivity before saving.
+Himanshu showed how we're working on anomaly-based alerting using Z-score detection. Instead of a fixed threshold, you set how many standard deviations from the expected mean should trigger an alert. The seasonality window (hourly or daily) controls how the baseline is computed, and the alert editor highlights the windows that would have fired so you can tune sensitivity before saving.
 
-This addresses a common problem with threshold alerts: users often don't know what numeric value to set. Seeing the potential alert windows update live as you adjust the Z-score makes it practical to find a threshold that catches genuine spikes without producing excessive noise. Alerts can be scoped to fire only when values exceed the expected range (not when they drop below it), and an occurrence setting lets you require the condition to hold for multiple consecutive data points before firing. Currently uses a standard moving average; exponential moving average support is planned.
+This addresses a common problem with threshold alerts: users often don't know what numeric value to set. Seeing the potential alert windows update live as you adjust the Z-score makes it practical to find a threshold that catches genuine spikes without producing excessive noise. Alerts can be scoped to fire only when values exceed the expected range (not when they drop below it), and an occurrence setting lets you require the condition to hold for multiple consecutive data points before firing. This is a work in progress, with the baseline computed as a moving average currently; exponential moving average support is planned.
 
 ## Notebooks generated by Claude {#notebooks-generated-by-claude}