seasonalityAnomalySensor v1.0.1

Seasonality Anomaly Sensor
For a given metric, it goes back in time and compares the current series value to the past series values.
seasonalityType: This setting controls whether the search goes back for the same time slots X (which is slots input) days back, or X weekdays back (e.g. go back X Mondays, Tuesdays etc…), depending on the day in a week in which sensor executes.
Default is weekDay. If you put days, it will just go back in time, for the past X days.
percentageFlag is by default true, meaning the boundary (change) will be taken into account as the percentage change between current slot and the past data.
slots: default value is 7
pastMinutes: default value is 60 minutes.
aggregateType: default value is mean.
boundary: by default 10

Aggregation types:

  • mean
  • medium

  • min
  • max
  • sum
  • count

Required properties:

  • resource
  • metric
  • aggregateType
  • pastMinutes
  • seasonalityType
  • slots
  • boundary
  • percentageFlag

External Settings:

  • waylay_data

States:

  • In Range
  • Above
  • Below

Raw data:

  • totalSum (double)
  • sum (double)
  • delta (object)
  • qs (double)