How do you assign alarm severities for variance, trend, flatline anomalies?
For variance changes, trend changes and flatlines, you can only specify whether you want alarms or not by selecting the checkbox in front. When selecting this option, DataMiner will decide whether the change is significant or not and what severity it has based on the past behavior of the parameter. This corresponds to the 'smart' option for level shifts or outlier.
If you would like to have more detailed options also for variance changes, trend changes and/or flatlines, feel free to post a feature suggestion at https://community.dataminer.services/feature-suggestions/.
Yes, indeed. At the risk of spoiling the exercise, the difference in variance is big enough so that critical severity is set automatically. If it weren’t there would currently not be a way make it critical.
Hi Jorge, more information on this topic can be found here: Configuring anomaly detection alarms for specific parameters | DataMiner Docs.
Thank you Tobe. Step 7 in the Anomaly Detection Tutorial asks to configure the alarm template to show a critical alarm at the start of the period with higher fluctuations, so I thought there was a way to control that with the variance settings.