We see a problem with "schedule refresh" that seems to happend after upgrading to oct 2020 (first version).
Some "larger" reports consume all the memory on the server and the server (service) crashes.
Prior to the upgrade have have not seen this problem.
The server has 32 GB of RAM memory.
We can see that it seems to be the Microsoft Mashup Container processes that consumes the memory.
In this case it starts 22 instances of this process. Looking in "SysInternals" Process Explorer we can see that each process consumes over 1 GB of memory (private bytes) however the "working set" is only ~262MB per process.
Looking in Task Manager we can see that the system uses "all" the 32 GB of the server.