First up, we’re not really going to change any configuration, we’re going to use Tableau data extracts to store the data and incrementally refresh the information.
It’s that simple.
Http_requests is essentially a transaction log with an id that increment for each new record, an incremental extract tells tableau to grab any records are have a larger id than it already has. Incremental extracts only ever add records so when the data clean-up deletes data in the source table they are not deleted from our extract.
Desktop_reporting we’re getting a little tricky as tableau repository maintains one record per product key, machine and action combination (or something like that) and updates these records with the latest information rather than adding new records. So this time we’ll tell the increment to grab any records that have changed since our last increment and add them as new rows to our extract. We run this increment on Last Report Date and if we run it daily you’ll be able to get some daily stats on your desktop (and reader*) users
(if someone explains the above in a nicer way I’ll edit it in)
I’m going to assume a general level of knowledge around tableau desktop server and publishing.
*You know Tableau Reader will report back to your Tableau Server as well? Just get the same registry keys deployed.
What you’ll need
- Access to Tableau’s Server repository https://onlinehelp.tableau.com/current/server/en-us/perf_collect_server_repo.htm
- I’m assuming you have Tableau server and desktop and
sooo does it work?
I've got over one year of http_request information, so determining performance of server / workbooks etc is a breeze
I only turned on the desktop_reporting one recently
Who's using desktop daily and how many users per day.