Context: existing Splunk installation I'm working with is not very robust when handling search requests due to sheer volume of searchable events.
The question here is - is there a way to make splunk disregard default sorting behavior and return first N found matches as quickly as possible?
The goal here is to use this in conjunction with head clause to make search return first matches as quickly as possible - it is totally OK if events would be presented without prior by-time sorting.
So the expectation is that this approach should make search near-instantaneous provided that filtering expression is broad enough and first N matches could be found very quickly.
Map/Reduce, the "sorting behavior" goes to the core of Splunk and cannot be disabled.
The best ways to address this problem are: 1) make the Splunk environment better match the search volume (add hardware or reduce the number of searches or both); and 2) make the searches more efficient so they return data faster.