Machine Learning Analytics Job error

Hi everyone,

I have a question concerning Machine Learning Analytics Jobs. I have already created an auditbeat index and data in it. Created also Machine Learning Jobs which I can browse within the Anomaly Detection. We are running the elastic stack in version 7.7.0
But I want to preprocess my data with Analytic Jobs and here is where my problem begins. I tried all sorts( outlier detection, regression, classification ) and I get always an error message. I dont understand the error so therefore I dont know where to search.
Hopefully someone had this problem before or has a tip. I receive the following error messages on the three choices of Analytics.

outlier detection -> Bad Request: [status_exception] Unable to start ml-outlier-test2 as no documents in the source indices [auditbeat-*] contained all the fields selected for analysis. If you are relying on automatic field selection then there are currently mapped fields that do not exist in any indexed documents, and you will have to switch to explicit field selection and include only fields that exist in indexed documents.

regression -> Bad Request: [illegal_argument_exception] cannot apply boolean mapping to field [aws.cloudtrail.console_login.additional_eventdata.mfa_used]

classification -> Bad Request: [illegal_argument_exception] cannot apply boolean mapping to field [file.setgid]

Thankful in Advance

1 Like

Data Frame Analytic jobs expect an entity-centric index, not a time-series index (like Anomaly Detection expects). By default, indices created by beats are time-series.

You can transform a time-series index to an entity-centric index using Transforms.

I think you need to describe what you want to accomplish more succinctly and then we can assess if Anomaly Detection or Data Frame Analytics is the right approach for what you want to do

1 Like

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.