anomalySensor v1.5.1

by thomas

searches for anomalies back in time (sliding window).

Properties
  • resource
  • metric name of the metric configuration
  • window number of seconds to look back for anomalies (default 3600). Alternatively you can use an iso8601 duration expression such as ‘PT2H’
  • consecutive number of consecutive anomalous observation that trigger an anomaly condition. If e.g. 3, then 3 observations in a row need to be anomalous before the Anomaly condition is triggered
  • params overrides for the analytics option configuration (see below)
States
  • Anomaly if there where at least consecutive anomalies in a row within the window
  • No Anomaly otherwise
Raw data returned
  • anomalieslist of all anomalies in the window (including those below the consecutive threshold)
  • last_observationlast observation in the window (whether it is an anomaly or not)
  • last_anomaly last anomaly that satified the consecutive threshold

this functionality requires an active TS analytics license for your Waylay tenant

Before using this sensor, configure the analytics options for this metric using the analytics dashboard. Alternatively you can give option in the params section that will override any options stored. See apidoc on how to specify options.

Required properties:

  • resource
  • metric
  • window
  • consecutive
  • params

External Settings:

  • waylay_analytics

States:

  • Anomaly
  • No Anomaly

Raw data:

  • anomalies (object[])
  • last_anomaly (object)
  • last_observation (object)