Open cmk1523 opened 4 years ago
Pinging @elastic/ml-ui (:ml)
https://github.com/elastic/kibana/pull/58152 has just been added to allow follow up files to be uploaded to the same index using Filebeat. This would supply some of the functionality asked for here, albeit in a way that isn't quite as convenient. It's worth noting that the file data visualizer feature is not intended as a replacement for the normal bulk ingest processes, but as convenient way to get some data quickly into es. With this new PR, the index and ingest pipeline can be easily reused for ingesting the rest of the users data.
In the machine learning section of Kibana, I'd like to see support for importing of multiple files not just one.
The use case: Any given kibana user wants to upload multiple similarly structured json files in order to not have to import each one individually. Kibana should treat multiple files as one if possible (e.g. files that have multiple lines of json documents). That should be possible for CSV data too. If columns or fields don't match, they can be detected as "new" fields within the machine learning system.