Analysis Workspace: User Settings (e.g. setting default options)



We've consolidated all threads related to setting default options into this thread, including:


Original ask for Anomaly detection on/off:

It's fantastic to have gotten anomaly detection in Workspace in today's release. One point of feedback however would be that the function should be off by default when adding a new metric, with an option to turn it on.

The reason we believe this is, workspace is used by a variety of end users throughout our organization (from technical to non-technical). For some of the non-technical users seeing an exclamation mark or anomaly has the potential to cause confusion and spurious concern. We therefore believe that having the default as off, would allow the anomaly detection to be more appropriately managed and used within our organization.

Thanks in advance,


17 Comments (17 New)



Second this!




Thanks for the feedback - one option that you can use to prevent unwanted users from seeing anomaly data is to turn off their anomaly detection permission in the Admin Console.  Then they will not see anomalies in Workspace at all.



On top of having the default set, when making new functions such as this, they shouldn't be automatically added to existing reports.


I have a few massive reports that I export for monthly comparisons, since the reports are slow to begin with, and the anomaly detection makes them slower, I have to spend 1.5 hours removing them all from each report.


Maybe there could also be a bulk add/remove function for all metrics columns?



Agreed--especially if this can be selected at the project and/or panel level, as jdungan suggests.



Please do this one soon. Not only does it bog everything down, but I'm sorry to say it, but the anomaly detection isn't very good.



I second this as well. The forced anomaly detection leads to lots of lag when trying to quickly compare metrics. While it's sometimes very useful, it turns up false positives often. An option to turn it off would be great!



The performance of the anomaly detection is actually rather problematic. Especially when it still bogs down the loading time on all reports, despite it being turned "off"....


If it's turned off, we shouldn't be stuck waiting for it to "check for anomalies".... why does it need to check when we aren't going to be using it? It just makes the reports slower for no reason.




Hi jdungan - thanks for the feedback.  It sounds like you may have encountered a bug (it shouldn't be checking for anomalies if you have it turned off).  Can you please report it through customer care and we'll get it addressed?  Thanks.



Hey what's the status on this? At 2.5 years old this idea is not new anymore. Either fix Workspace performance, or make turn this feature off by default, or at least add an setting to a Workspaces where I can turn it off for the whole workspace at once. This really drags down performance, and it's really a pain to have to wait for anomaly detection to finish loading every time I add a line graph... I love AA but man the performance just kills me.




Thanks for the feedback - a few updates/notes on this.  We are actively working on providing a bunch of customizable user defaults for Workspace so you could potentially turn off anomaly detection for new projects (along with a bunch of other settings).  We know this functionality is important to you, and are doing our best to prioritize it on the roadmap.

I do have a follow-up question though, just so I'm understanding better.  For time-series data (ie, data shown on a line or area visualization), we make two separate requests - one for the normal data, which normally returns very quickly, and then a second request for the anomalies (when it's saying "searching for anomalies..."), which usually takes a little longer.  We intentionally split the requests so that it wouldn't slow down the initial request.  So I'm curious as to why you feel like this really drags down performance, if it doesn't affect the initial request (it's just extra, and shows up when it's done).  Thanks for your help.