I am running a desktop flow called from a cloud flow, which extracts data from a web page, scrolls the web page (it uses page-down keys) and then re-extracts. This is running in unattended mode on a virtual machine. What is very strange is that this works well when I manually trigger the cloud flow (still running unattended) but not when it runs the schedule. I have it logging how many rows it retrieves, which tells me it is often not successfully scrolling down when it runs on the schedule. (Without scrolling, the table initially retrieves the first 100 rows, so if exactly 100 rows are retrieved, I know it is not scrolling).
The workflow:
- Load the web page
- Focuses the window
- Sets focus on the text field on the web page which houses the table (the table has a scrollbar but is not in an iframe).
- Send keys ({Next}{Next}{Next}{Next}{Next}).
- Extract data from web page.
I am sharing screenshots. I cannot share the web page because it is a private database with sensitive information. Again, when I click "run" on the cloud flow, it seems to work fine. But when I runs on the schedule, it's mostly not scrolling down.
In the screenshot of the log, the highlighted row with the higher number of rows was a successful run where I triggered the cloud flow.

