You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the feature is limited to only support single-stream detectors to be both (1) created and (2) associated to visualizations. This is due to added complexity regarding automatic high-cardinality-detector generation based on visualization configurations, which need to be narrowly specified and have a terms aggregation on the x-axis.
However, even given that niche use-case, a simpler and more feasible iterative approach would be to allow existing high-cardinality anomaly detectors to be associated to visualizations, and be overlaid on top of visualizations the same as single-stream detector anomalies. Automatic detector creation logic can be added later.
Some technical things to consider:
is special rendering logic needed in the associated detectors or view events flyout to support HCAD alongside single-stream?
what added logic is needed to fetch and format high-cardinality anomalies and include them in the chart generation?
The text was updated successfully, but these errors were encountered:
Currently, the feature is limited to only support single-stream detectors to be both (1) created and (2) associated to visualizations. This is due to added complexity regarding automatic high-cardinality-detector generation based on visualization configurations, which need to be narrowly specified and have a terms aggregation on the x-axis.
However, even given that niche use-case, a simpler and more feasible iterative approach would be to allow existing high-cardinality anomaly detectors to be associated to visualizations, and be overlaid on top of visualizations the same as single-stream detector anomalies. Automatic detector creation logic can be added later.
Some technical things to consider:
The text was updated successfully, but these errors were encountered: