Access Denied Error While Using Send Global Key While Running From Scheduler

[devneetmohanty07]

Hi All,

One connection of mine recently faced an error while trying to use the Send Global Key action from her navigate stage. The error described - Access Denied.

While running the process from Control Room or from Studio, this error does not appears but while running the same from Scheduler, she is getting this error. I tried looking up for something in the community and saw that RDP settings might need to be changed somewhere, however it was not clearly mentioned which setting if there is one.

Can someone please help to go through some possible troubleshooting steps or navigate me to some post if I missed by any chance?

------------------------------
Regards,
Devneet Mohanty
Intelligent Automation Consultant
Blueprism 6x Certified Professional
Website: https://devneet.github.io/
Email: devneetmohanty07@gmail.com
------------------------------

[NicholasZejdlik]

RDP will conflict with any surface automation techniques, including the Global Send Key action. It will also affect elements spied out with active accessibility mode. See this article and the remote access tools guide. The issue is how the sessions are handled. Blue Prism generally needs to be run on the "console" session in order to function properly, RDP causes problems with that. This is why you are able to run it from studio OK, but it fails when running unattended on scheduler.

------------------------------
Nicholas Zejdlik
RPA Developer
------------------------------

[devneetmohanty07]

Hi Nicholas,

Thanks for your reply, it seems that there is no surface automation being performed in the process. The write stage is not working on the website so keystrokes are being passed as an alternative, Also, the Blue Prism Runtime Resource and Client is within the RDP environment itself but yes I knew RDP is not suitable for deployments however, it would require approval of an infrastructure change around the client end which I am not sure how much feasible it would be at this point.

Main intent for posting the thread is to know is there any alternative way that this issue can be resolved since getting Access Denied error on just one stage whereas the rest of the stages working fine seems bit strange to me but again like your post mentions it might be an issue with how Blue Prism handles the session.

------------------------------
Regards,
Devneet Mohanty
Intelligent Automation Consultant
Blueprism 6x Certified Professional
Website: https://devneet.github.io/
Email: devneetmohanty07@gmail.com
------------------------------

[NicholasZejdlik · Accepted solution]

Win32 and HTML elements do not need an active session, so if the process uses just those elements up until the Global Send Keys stage, that could explain why it is only failing at that point. You won't be able to make send keys work without the active session - that's on Window's end - but since it's a web page you're working with, you may be able to write the text differently. For example, you might be able to insert some javascript and invoke a function that will fill in the problem element. Alternatively, I've had a lot of luck bypassing some pages altogether by manually POSTing form data to a URL instead of using the UI, although that tends to be more complicated.

------------------------------
Nicholas Zejdlik
RPA Developer
------------------------------

[AlexanderBopp]

I was facing the same problem using Send Global Keys on a Javascript based HTML applciation. What helped for me is to add an element in Application modeller which identifies the whole browser window (Win32 spy mode). Then in my Navigate stage I could select Action "Activate Application" for this element and also "Maximise" to ensure it is not out of screen range. After this my Send Global Keys was working.

Might be a workaround solution for someone - hope it helps!

------------------------------
Alexander Bopp
Business Automation Engineer
Credit Suisse AG
Europe/Zurich
------------------------------

[karan.khanna.gpi]

Can it work if we make some changes in the login agent to make the screen active. How is it possible that UI mode doesn't work with Chrome automation while running from Control room and works while running it from Studio.