We are trying to index long JSON files. Each JSON file is one event.
As performance is more important to us than storage space, we wish to extract all JSON fields at index-time.
We tried to use the INDEXED_EXTRACTIONS=JSON configuration, but it seems that it does not extract all the available JSON fields
(For example, there are many fields missing from the "Interesting Fields" section).
We conducted some tests, and were successful in extracting ALL fields using KV_MODE=JSON , but it's no good for us because of the aforementioned performance issues.
How can we make Splunk indexer extract all fields during index-time? Are we missing some kind of configuration regarding the depth of the JSON?
Just because fields are not in the Interesting Fields does not mean they have not been extracted; it just means that they are not "interesting"! You need to click on the All Fields link to pull up the fields popup and search for the "missing" fields there. In all likelihood, they are there (not really missing).