I would like to queue up a large number of unique searches (approximately 500,000) to be run from the command line in a queue. Right now, they are going one at a time from a shell script and they're painfully slow. Is there any way to stick searches into the jobs list and retrieve their data later? Alternately, is there any other way to run massive numbers of searches in a batch?
The REST API lets you start a search and returns a job ID for later retrieval of the results, that way you could run several searches concurrently. I have not yet tried firing a large number of searches at once though, ymmv.
http://docs.splunk.com/Documentation/Splunk/5.0/RESTAPI/RESTsearches