At this point in our development, we have about 12 distinct alerts that we've created for ourselves. I thought I'd provide a few examples in case anyone here has wanted the same, or needs some detail around how to implement them. These are roughly in order of popularity:
Some Alerts We've Built:
Extract Refresh Failures
This is currently implemented as a Simple Alert that I've allowed individual Tableau Server users opt into by subscribing on an hourly Alerts schedule. The data is user-filtered so that they will only ever see their own extract refresh failures. I use a customized view for Admins to subscribe to that filters out some of the noisier extracts that frequently fail intermittently.
My Nirvana for this alert is one that requires no opt-in on the users's side and just emails them the failure. The only thing stopping us from building it is that we have failures for lots of reasons--server restarts, databases going down, or again, internal issues that resolve themselves the next time around and never get noticed.
Extract Refresh Delays
Another alert focused on extracts, this one looks at all extracts associated with a high-priority Schedule and alerts if any of them become more than 120 minutes stale (the schedule is expected to run every 15 minutes). This is for us Admins only and guards against human error if the schedule should be accidentally left disabled for any reason.
"Critical Viz" Alert
We have a process that runs through and renders our most critical vizzes, and pushes data to a PostgreSQL database. This alert fires if more than 5% of our attempts fail within an hour, which would likely indicate a stability issue with Tableau Server.
Change Management Approval
Our security guy built a daily alert that emails him should someone inappropriately close a Change Management ticket without it having been through the proper approval process. Several others have subscribed to this one as well.
A user reported to me that they had noticed stale data in one of the data sources we use monitor Support Cases. After looking into it with the ETL team, we couldn't pinpoint the issue--everything looked fine and the issue disappeared. I built an alert that triggered based on the published data source being stale, but embedded it into a dashboard that pulled dates from all the upstream ETL systems for comparison. Next time it occurred, we were able to determine that it was an upstream issue between two of the SQL servers.
There are still a lot of alerts we want and need to build! And this is just from my own head, forget about the rest of our enterprise...
Alerts We Still Want:
These are all still just stuff I want....
I want to inform people when they have Workbooks or Datasources that haven't been looked at or touched in a long time, and bug them to ditch 'em. Probably on a quarterly basis. For users who are unlicensed, but still own content, I want it to send the email to their Project Leader and let them decide on what to do with it. Will need some usage stats embedded in the email to aid decision-making.
Subscriptions on Tableau Server have a tendency to go out late when lots of people subscribe to long-running views or worse, entire workbooks. Rather than check our viz for this every day, we want to simply be told that some subscriptions are going out too late so we can investigate.
Best Publishing Practices
I'd love a way to email users who publish workbooks that have one or more implementation details that don't conform to Server best practices. We'd need to obtain that data somewhere, of course, but that's the trickiest part.
Welcome Email for New Tableau Users
When a user logs in to Tableau Server for the first time at our company, send them a welcome email with some basic resources. When they first publish to Tableau Server, send another email with tips, tricks, and best practices. When one of their views gets over N views per day, send them (and the Admins) another email with a congrats and some resources on scaling.
What kind of alerts do you want to set up?