Questons about Compound DFE

#1

For Compound DFE, the analytic would only be invoked when the thresholds for all the time series in the DFE are satisfied.

I have a few questions,

  1. what is threshold for a time series? Is it data completeness? (i.e., how much of potential data has been collected in the period)

  2. If not specified, what is default threshold for a time series?

  3. if a Compound DFE contains two time series. Metric of one time series is ElectricityConsumptionInterval, with expression sum(measurements.(resource.id == 'electricity').normalized.data.quantity)), and metric of the other time series is NaturalGasConsumptionInterval, with expression sum(measurements.(resource.id == 'naturalgas').normalized.data.quantity)).
    Does the Analytic Engine distinguish electricity measurement and natural gas measurement when checking thresholds? If not, does that mean as long as enough (greater than threshold) measurement data are collected, both time series will be satisfied, right?

  4. for a time series, if its metric doesn’t contain any measurement nor event (as following shows), how could we know if the threshold is satisfied or not? Always satisfied?

{
  "id" : "AnomalousUsageThreshold_Facility",
  "name" : "AnomalousUsageThreshold",
  "srcType" : "Facility",
  "expression" : "identity(0.85)",
  "description": "Default threshold for anomalous usage alerts"
}
0 Likes

#2

threshold has been removed as a feature starting server v7. The analytic will always be invoked with the results of the metric evaluation. It is upto the author of the analytic to decide if the available data is useful for the analytic or not.

0 Likes