I think you need to define your goals much more precisely, but lguinn is on the right track. Making the general request of "How do I use X with Y?" when both X and Y have a myriad of possible uses is difficult to answer. As an example, "How do I use flour with eggs?" Well, are you wanting to make a cake, some bread, a batter, pie crust, pancakes, waffles, or what?
Let's assume you have some R code that you wish to run against data stored in Splunk. Depending on how you want to "integrate" the two, this could mean one of (at least) two paths.
- You could use R programs as custom search commands in Splunk. There would be some effort here to bridge the gap between Splunk's supported Python/Perl search command interfaces and your R code. But, it's a bridgeable gap. Your R program would receive (essentially) csv on input, do whatever it is you want it to do, and emit csv on output. The Splunkweb GUI would still be present, and your commands written using R could be intermixed with existing Splunk search operations. These results of these could be placed on dashboard panels, used in alerts, etc...
- You could make a standalone R program that uses Splunk data via an existing Splunk API. A quick (10 minute) reading of the R documentation and a googling on "R REST" shows that R has some API's already for accessing remote REST endpoints and interpreting XML coming back from them. This may give you a leaping-off point to writing a Splunk SDK for R - but it would be a nontrivial task. This would be using Splunk almost exclusively as a datastore - you would need to provide your own visualizations and user interface.
