yesterday
yesterday - last edited yesterday
The only behavior I have seen that mimics your scenario is if you are opening the Blue Prism UI when inside the remote session. I had this issue specifically in 7.5. The same thing did not happen in v6 for me. What seemed to be happening was that even if the runtime resource process was running beforehand, when opening the UI (in our case through vCenter/vSphere) and closing the UI, Blue Prism would stop the runtime resource process just like it would when in development such that it closes the private runtime resource process. In this case of course it should not close the process down because it was running beforehand already.
If, by chance, this is the same issue you are having, then you have two options that I know of.
1 - Upgrade to 7.5.1 which does not have the issue, at least as far as I've experienced.
2 - Regardless of version, before you exit the session, you (1) close the UI, then (2) you check for and close the runtime resource process if it is running still, then (3) you go to the Startup folder and re-run the runtime resource startup shortcut/batch file or whatever, and then (4) exit the session. This should result in the runtime resource "staying" connected after exiting the session. This of course assumes some things about SCCM which I've never directly used before, so it could be different.
P.S.- There are other reasons to not stay on 7.5 as well though. For example, I believe on that version the wait stage functionality was changed but then was reverted as of 7.5.1. The timeout behavior was changed so that it acts like a max time for blue prism to find an element. If Blue Prism takes 6 seconds to find an element and you have a timeout set to 5 seconds, then it will fail as if the element doesn't exist even if the element existed before starting to check. This was a big enough deal for us that we couldn't use 7.5, specifically because we like to rely on a 0 second timeout on wait stages sometimes which causes (in versions outside of 7.5) Blue Prism to check over the DOM only once for an element and move on; in 7.5 this doesn't work because it hits the 0 second timeout immediately and goes on the timeout path even if the element exists.