outlier Metrics Operator
outlier operator identifies metrics data points that are outside the range of expected values. Outliers help you spot unusual behavior in your metrics visualizations and track the behavior over time.
outlier tracks the moving average and standard deviation of a time series over a specified time window, and calculates a threshold band, outside of which data points are considered outliers. You can use optional qualifiers to specify the time window, the number of standard deviations beyond which a data point is considered an outlier, and the directionality of the deviation.
At this time, using the
outlier operator in a metric monitor is not supported.
metric query | outlier [window=<#>, threshold=<#>, direction=[ +- | + | - ]]
windowis the range over which to calculate the moving average and standard deviation of the time series.
windowcan be specified with time units (s, m, h), or it can be specified without time units. Default: 5m.note
If you use
outlierin the Classic Metrics UI, if you specify the
windowparameter without supplying a unit of time, the window duration applied will be in the units used in the quantization of the query.
thresholdis the number of standard deviations from the moving average that defines the threshold band. Default: 3
directionspecifies what deviation direction should trigger violations: positive deviations (+), negative deviations (-), or both (+-). Default: +-
In the visualization, the threshold band is the part shaded in pink. The outlier values are represented by the pink triangles.