We have a simple use case to alarm when the throughput drops. However, on this environment, the data is very spiky, ranging from 0 to hundreds of Mbps many times within a minute. On this case, we need to raise an alarm if the "average throughput" drops below a statically definable threshold. Below an example of the calculated averaged trending and the respective alarm colouring
I believe this cannot be done with hysteresis because if the average drops we don't care about a spike high, it is just a matter of applying the monitoring to the calculated averaged trending datapoints instead of the real-time data display datapoint, assuming the calculation is across a 5 mins timespan.
Is there a solution by just configuring it?
Hi Elvio,
Good question and nice use case!
I believe the only option today is to add a new parameter (or column to the table) with e.g. a moving average based on the data you have (like the blue line in your drawing). In other words, this would require a change in the driver. And then you can do the alarming on that new parameter.
Bert
Just to further complement on this Elvio. While it is not possible today it will be supported through GQI. Today you can use GQI already to display these use cases on a dashboard (https://community.dataminer.services/video/analysing-historical-trends-with-dataminer-gqi/). Of course this does not give you any 24/7 alarming yet. But we are working on a solution to run GQI queries 24/7 in the background, in which case you will be able to alarm on them. But that’s not yet available.