@Damien O'Connor Thanks for the response back. I am using the Emma api (email platform) https://api.myemma.com/api/external/mailings.html. The thing that seems different is that the paging is not passing back any information on the paging in the body of the response. Instead, you get a max count from an api call using a count query parameter, and then construct your paging in 500 entry chunks, using a start and end query parameter. The copy activity paging rules seemed to support this fine, by putting in a parameter and selecting the range and offset. However, it never seemed to recognize the parameters properly. I tried both using curly braces and without, but neither worked. Any thoughts would be greatly appreciated I agree, the for loop is probably not an accurate test of running parallel - and I do control the parallel in the notebook to prevent the throttling, but am still interested in seeing if I can get the pipeline to work, with the pagination rules, as I think it could be much faster. Here is a screenshot of my last attempt. Not sure if I should be using the absolute url instead - but was not sure what that would look like.

Hey - thought I would give an update here. I actually go the paging in a pipeline to work if I just set one parameter with a range and then calculated the end parameter. Somehow, setting two parameters, the start and end, each with a range, seemed to create the fatal issue. However, I don't think this does parallel calls very well, and it was in the end, much slower that using async, semaphore architecture in a notebook.

@René Tachon Are you using a default semantic model created within Fabric, or a semantic model that is created using the PowerBi desktop. I have found they can behave differently and found times where only a semantic model created from PowerBI desktop had all the functionality I needed.

@Antony Catella interesting. I have not tried to customize SharePoint in a while. It was tricky back then. Thanks. Will be checking that out.

@Todd Magnusen You need to create a Service Principal and then write the code to get the token for the authentication, as that solution kind of walks you through. I am not trying to have a filter based on login though, which is a little tricky, because at least in my solution, I am using a single account to do the login. However, you can do embedded where you pass back the individual's account too, but then you are back to a different licensing model. @Diana Geyer was also trying to do filtering based on customer logins. Diana - did you get anywhere with that?

Yes, I have a pipeline that just orchestrates the running of everything else, so that I can have any flavor I need. I also put in error handling so that I get an email if anything fails.

We are a small organization using F8 for Fabric and PowerBI. One way we were able to afford this was to no longer need to have PPU's for all those who needed to view reports. Instead, we are using the embed token - embed for your customers (https://learn.microsoft.com/en-us/javascript/api/overview/powerbi/embedding-solutions) and displaying the reports within our CRM system. Seemed to be the best way at the time, but curious if that is still the best option.

@Mahmo Ud We are using F8 - and have had capacity issues running semantic layers that are too resource intensive. I have noticed issues when running a notebook - and have had to go to the monitor and stop other notebooks that have not time out yet, in order to get that one to run. Have you checked the monitor to see if something else is still running?