"Resume after delay" gets cancelled in Orchestrator after X hours


I have built a robot that runs a process successfully, and because the result is not visible before the next day I am using the “Resume after delay”-activity (part of UiPath.Persistence.Activitites) to continue the job when the system have made necessary changes to be able to continue.

I have successfully tested this functionality on a different robot with a delay of 5 minutes, but when I use it on this robot and set the delay to +1 day the job got cancelled after approximately 12 hours without any trace in the logs. Last info in the log is the expected “ececution suspended”-message and nothing that could explain the cancellation.

This happened early this morning, and I was expecting it to still be suspended when I checked.

Can anyone explain this behaviour?
Is this caused by the Orchestrator (community license is used during developing), or are there a max delay on this functionality before all suspended jobs are cancelled?

The same happened again during the night. It got cancelled at a different time this time.
What is going on here?

We still got issues with the «Resume after delay»-feature and need to figure this out.
I have now made the easiest kind of test to prove that it`s nothing in our environment that can cause this. How can we make this work?

In the Orchestrator we see the following (all 4 tests started shortly after each other):

The only difference between these 4 tests are the time of delay.

2 hours test successful log:

Logs from a failed one (they are identical):

As you can see nothing else happens in the log after it`s suspended.

Workflow is very simple and just adds a value provided before the delay with a value provided after the delay. There is no interaction with any web browser or external systems besides the UiPath components.

This is how the workflow looks like: