@souenzzo I wouldn’t normally expect that number of facts to be much of a problem
however, you can certainly write rules in a way that cause a problem (for smaller numbers even)
@souenzzo what do you mean by “strange exceptions”? In terms of number of facts, the only real limit in Clara should be the memory available.. but of course if you have problematic rules (not necessarily incorrectly written, just poorly performing) you can run into problems. If you can find a reproducing case that seems like it performs more poorly than it should feel free to log an issue. We’ve mostly optimized Clara against the examples we have at Cerner but representative benchmarks of other workloads could be interesting/useful.