Since I updated Kibana and ES to 6.0, my visualizations that relies on aggregations do not work anymore.
I see a message saying "No results displayed because all values equal 0." instead. If I select a time range that goes before the migrations they work perfectly fine.
It happens in every visualizations that rely on aggregations by keyword terms. But it works still fine when they use the string format (for example I have graphs with filters like intent:"something-*")
The new documents are under the same index as the old one
Yes
Yes
One thing I am not sure about and don't know how to check, is if the new documents have their keyword field created properly.
I don't really understand how keywords work yet, and I don't know if the keyword value of a string field is stored in elasticsearch or if it is just an abstract notion.
If the documents are indexed into the same index, the type of the intent.keyword field is bound to be the same. Would it be possible for you to show me the mapping of that index (by running GET indexname/_mapping in the Kibana devtools with "indexname" replaced by the proper name)?
Oh sorry @weltenwort I figured it out myself!
Indeed the new documents have _type:doc while the old ones have type _type:logs.
I now understand that the issue is that the mapping for type doc does not have keyword subfields.
Am I correct if I say that sending the following request:
Thanks for sharing your solution. As to why the document type has changed, it is probably a change in your ingestion pipeline (beats, logstash, ...?). The removal of mapping types I linked to might be the underlying reason.
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.