I am trying to create a forecast job via console and when I set expires_in to "1m", it creates the job ("acknowledged": true) but then it doesn't expire after 1 minute.
The documentation redirect us to Time units which means that we can use d, h, m, s, ms, micros, nanos.
Is it a bug, or is there any other way to do that?
POST _xpack/ml/anomaly_detectors/example/_forecast {
"duration": "56d",
"expires_in": "1m" }
ML has a daily maintenance task that does things like clean up expired forecasts. Since it only runs once per day, setting expires_in to a value less than 1d doesn't force that maintenance task to run any quicker than it normally does.
Thank you for your response!
So what time does this daily maintenance happen?
For example I set another forecast 13/8/2018 at 10:00am with "expires_in": "1d" and it hasn't been deleted(14/8/2018 - 10:30am).
Daily maintenance runs at a random time between 0:00 and 2:00 am so the forecast you created on the 13th won't be deleted until 0:00 - 2:00 am on the 15th when it is more than 24 hours old
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.