Especially when interfacing with external systems like Grafana. This would be a nice, native way within Graylog to specify the appropriate datatype without mucking with template modifications on a per-index-set basis within Elasticsearch. I was honestly a bit shocked that I saw this…
Expected geo_point type on field [src_coordinates], but got [keyword]
I understand why Graylog isn’t storing it in that format from the lookup_value call natively, but currently there is no function to specify the appropriate type in the pipeline rules.
@jochen@lennart I explicitly stated above that this would be to avoid doing exactly that, modifying index templates, which again must be done on a per-index-set basis.