Adding ML charts to a dashboard

[liorg2]

Hi
Is it possible to add ML charts (with anomalies and bounds) to a custom dashboard?

[Peter_Harverson (Pete Harverson)]

Hi,

The charts and visuals used in the ML Anomaly Explorer and Single Metric Viewer are unfortunately not currently available as Kibana visualizations. The ability to add the anomaly charts and swimlane to Kibana dashboards, as Kibana embeddables, is something we are looking at for a future release.

In the meantime, you can use out-the-box Kibana visualizations to build something similar. For example, you can use the Time Series Visual Builder (TSVB) visualization to overlay anomalies on top of metrics through its annotations feature, as demonstrated in this blog post. This ticket also provides some extra information on displaying machine learning information in Kibana dashboards.

For jobs with model plot enabled, the bounds in the Single Metric Viewer chart use the values of the model_upper and model_lower from the result_type: model_plot data of the job stored in the .ml-anomalies-shared (if using the default ML result index for the job).

Hope this helps,
Regards
Pete

[richcollier]

Also, perhaps relevant:

Show multi-metric machine learning job heatmap in Dashoard

Hello, Can we illustrate multi-metric ml job heatmap in Dashoard? I have read through the forum and I tried to create a heatmap or swimlane but I cannot see the values of anomaly score when I split the data (e.g. by product) in machine learning job. It shows only the data from job_id. Thank you very much in advance! /Angelos

And, if you want to know more about anomaly scoring, see the ever-useful blog Pete and I wrote:

Machine Learning Anomaly Scoring and Elasticsearch - How it Works

[liorg2]

thanks, it is relevant

[liorg2]

thanks very helpful