[rules-users] Drools and StatefulKnowledgeSession performance

Wolfgang Laun wolfgang.laun at gmail.com
Mon Dec 16 06:25:40 EST 2013


And you have made sure that you are not accumulating any objects, not
necessarily fact object.

Have you tried to see whether the effect occurs with a single rule, too?

Which Drools version are you using?

-W


On 16/12/2013, Peter Troelsen <peter.j.troelsen at gmail.com> wrote:
> Hi,
>
> I have implemented a small amount (around a 100) rules. Then I made some
> performance tests for my implementation to see how fast my implementation
> is. However I have noticed a tends whenever running tests for a long time.
> It seems that every time i insert an object, validate it, and then remove
> it from the ksession using .retract(t), the following  insertion gets
> slower and slower. This doesn't really show in the beginning, but after the
> first couple of million objects, i really starts to escalate..
>
> Have anyone noticed the same kind of thing? I am positive the i retract
> *all
> *my objects after validating, but still, it seems to get slower. Is the any
> way to do an extra clean, without retracting the rules or do I simply need
> to reinitialize the ksession occasionally?
>
> Thanks in advance!
>
> --
> Peter Troelsen
>


More information about the rules-users mailing list