Hi. Currently trying out X-pack. I successfully ran a multi-metric machine learning job for anomaly detection, and I can navigate through Anomaly Explorer and Single Metric Viewer just fine. However, I'm interested in looking at the visualizations for my non-anomalous data, but it seems that the Split variable dropdown menu in Single Metric Viewer only allows me to select those that are anomalous? Is there a way for me to see the time series visualizations for the data splits that don't have anomalies? I'd like to compare visualizations to get a better understanding of what a non-anomalous time series should look like for this data set.
This is true - the ML UI is currently not designed to explore non-anomalous data. The good thing is, there's an entire set of other tools in Kibana that will! Check out Timelion or the Time Series Visual Builder components of Kibana. Here's a recent TSVB blog that shows how to overlay anomalies on series you plot (if the anomalies exist): https://www.elastic.co/blog/time-series-annotations-and-anomalies-with-kibana
In this way, you can simultaneously plot anomalous or non-anomalous series (or both)
2 Likes
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.