Do downsampled queries change as time progresses?

I just received an alert on a metric going below a threshold. A few minutes later I got an [OK] email and the value was different. Please see the images for clarification but at 12:00 you can see in the alerting email the value is below the threshold but then later the value changes and I receive an [OK] email. The final photo shows the non-downsampled data and the data clearly has a massive spike.

I was under the impression that the 1h downsampled data wasn’t constantly updating, it was the average of the previous hour but it appears that is not the case?

Can you show the “state history” for the alert and the alert settings?

This could be a bug.

I’m attaching the state history and also the alert parameters in case it is something I set up incorrectly that I just have not noticed.


Here is the graph with all parameters as well.

Is this an issue? Wondering if I should submit an issue (I’m not sure how to do that honestly) or something else?

Another weird issue,

I downsampled the temperature tag (the “…_TT_8301_201.TPV1” tag) but yet the alert is still alerting based on the actual data it seems since the downsampled temperature value has not changed, it reads 954, but the most recent alert is based on a temp of 1077.

I don’t understand how these alerts are supposed to be working because they don’t seem to be working the way I had expected. I assumed the downsampled data would be the data that the alert bases on but it is clearly taking temperature values from the non-downsampled data.

Also including the alert rules: