Usage analytics is a set of analyses that receive information about user actions, or usage events, such as clicks or viewed items, on the SharePoint site. Usage analytics combines this information with information about crawled content from the Search analyses, and processes the information. Information about recommendations and usage events is added to the search index. Statistics on the different usage events is added to the search index and sent to the Analytics reporting database.
Came across this situation with another customer's deployment where Site usage Analytics reports don't work for one Web-application in the farm , while it worked fine on others .
For an overview of how Request Usage data is processed ,
1. Usage Events are written into .usage files that exist by default at C:\Program Files\Common Files\Microsoft Shared\Web Server Extensions\15\LOGS\RequestUsage
2. Microsoft SharePoint Foundation Usage Data Import - Timer job , updates the wss_logging database & processes data to update the so called Event store, which has folders for each day .
3. Event Store Folder exists on those servers which Run the Search 2013 Analytics Component & location as seen in my setup is at
"C:\Program Files\Microsoft Office Servers\15.0\Data\Office Server\Analytics_81c88675-044e-4c81-810e-9c8e001405cf\EventStore\20130705"
4. Microsoft SharePoint Foundation Usage Data Processing -Daily Timerjob which processes the Event store data & populates the Search Reporting database from where reports are pulled .
If you happen to see the entries in Event Store files , you will notice that all requests show up as Default Zone URL as that's how we store them there even if the end user request was to any other zone url of the Web-application. Which goes with the Guidance we have for crawling the "Default Zone" url of the web-application as Mentioned here
So the Reason this was not working was that we had the Intranet Zone URL in the search content source so the crawled items in index did not match the entries in Event Store & hence the Analytics processing would not get any output for that raw Dataset. All the Way one more reason to always crawl the Default Zone URL in search Content Source .
Here is another blog which explains other problems which may arise because of crawling Non-default Zone URL's .
POST By : Rajan Kapoor [MSFT]