When creating or editing existing Anomaly Detectors, sometimes they seem to get stuck in the “Initializing” state for more than a few days. I have noticed this a couple of times. When restarting the entire OpenSearch instance, the detector seems to move to a running state quickly (within a couple hours). It seems like it was stuck. Is there any way to debug what might be happening with the detector?
Also, can you provide an input on how long it takes for a detector to complete its training? Does that happen only in the Initializing state, or does that also happen after the detector moves to Running? Is there a way to turn up the logging level to see the progress of training?
Detector interval: 5 minutes
Detector delay: 1 minute
Feature: Sum
Uses categorical fields
Message rate: 35 messages/s