What happens is that you have a after start up microflow, which converts data from one type to another. Locally your data set works with this behavior, But on an other (cloud) node the data set contains data which doesn't match the used microflow expressions.
In other words; your after start up microflow isn't correctly modeled.
How to find and solve;
You might not find this easily, depending on the complexity of your microflow. So if its to complex, can do the following
For closure and in case this helps anyone else, the issue appears to have been in the else clause of the statement:
parseInteger(substring($ProcessingRecID, 0, length($ProcessingRecID) - 9)) * 1000000000 + parseInteger(substring($ProcessingRecID, length($ProcessingRecID) - 9, 9)
Changing instead to specifically group the multiplication operation to explicitly define precedence:
v v
parseInteger( (substring($ProcessingRecID, 0, length($ProcessingRecID) - 9)) * 1000000000) + parseInteger(substring($ProcessingRecID, length($ProcessingRecID) - 9, 9))
This of course didn’t change the operation, but now builds and runs on cloud as it did on local.