Currently I randomly sorting content based on a provided seed that ties with the users session cookie in order to increase content exposure on one of my services, this seed assures that pagination works appropriately for that client. This is effective, however I'd like to start tracking views for each piece of content in elasticsearch, then sort based on least views. This data collection is huge and with a random sort it takes along time for all content to eventually emerge, this collection can get as big as 5 million documents. It will be difficult to get all of this content to emerge, but we definitely need a better approach or most of our content is sitting unseen. It would also be a great metric to track, and would further allow us to better approach the issue.
What I'm wondering is:
hit
?