Can you try just creating the job even if Job Validation is returning the error? (Job Validation won't stop you from trying to still create the job even if it returns an error)
Let us know if that works. If yes, please also let us know if the results in Anomaly Explorer and Single Metric Viewer show up correctly.
In the meantime, we're trying to reproduce the issue with job validation you're experiencing.
As I understand, showing anomaly score reason is amount of data.
It means that when I tested first, used one day's sample but at used quite many date samples (e.g. 4 days)
Though I can see result of M/L API still show not valid Job validation when I create new job. At least now I give up why this happening. Below things are my test capture.
thank you for your reply again.
Thanks for all the details you provided. We identified a bug with job validation: It fails to identify the date field if it's a nested field, for example like in your case metadata.receivedAt. I filed a bug here: https://github.com/elastic/kibana/issues/24129
Update: A bugfix is in and will be released with the upcoming 6.5 Elastic Stack release.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.