Skip to content

Metric Graph Legend

When viewing a recent time range the metric is displayed at the frequency at which it was recorded for analysis.
At longer time ranges data is downsampled for performance reasons.

Metrics that increase monotonically (they only go up) are displayed in graphs in two ways. Their derivative value appears alongside their non-derivative value.

Data points are expected to fall in this range under normal circumstances. Data points that fall outside the expected range are anomalies. When a large time range is requested it's not performant to plot anomalies as yellow triangles. Instead a vertical gap in the expected range indicates where an anomaly has occurred.

Anomalies, represented as yellow triangles and vertical gaps the expected range, are unexpected data points. They are not similar to recent data and don't match a pattern of expected behaviour either.

Matches, represented by green triangles, are data points that matched with an expected pattern of behaviour and and such have been classified as normal by Anomify.

False positives are data-points which started as anomalies but were relabelled as false positives as part of the training process. They are also labelled with green triangles on the graph and can be selected.