Is DSV data adaptor type really needed?

Hi there

I’ve just upgraded to graylog-server 2.4 and noticed the Lookup API now supports DSV. From the looks of it, that could have been emulated by simply allowing URLs to define the path in the existing CSV format? ie support “http://xxxxxx/yyy.csv” as well as “/etc/graylog/…csv” (which is really file:///etc/graylog…csv")

(although the “ignore lines with #” would be needed too)

BTW, is there any timeline for allowing us to have one CSV file data adaptor that contains multi-values? Or even a multi-value JSON adaptor that was loaded once (between them all the work is there - just needs to be expressed differently?)

ie so that I can have a file (don’t care if it’s CSV or JSON) that contains multiple attributes per key, eg an asset database so that I can map a hostname to it’s OS, owner, hardware detail, etc


Fetching data from a remote location is inherently different from reading a local file (think timeouts, proxies, authentication, encryption, etc.), so in my opinion having two different lookup adapters for that is warranted.

Feel free to file a feature request at Issues · Graylog2/graylog2-server · GitHub if you think otherwise.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.