Is it normal for OpenRPA to consume that much memory after running a couple of simple workflows?
If I kill it, and reopen, it idles at about ~300MB RAM, but if I run couple of simple workflows (get elements, some for loops), it shoots to 1GB and above.
Also, I’ve read Allan you mentioned couple of times that the designer is memory leaky, and that in prod you run it without the designer, only as a service, so it shouldn’t be a probleme there. Are you referring here to HD robots? or is there another way to run it as a windows service?
There is no “runner” for OpenRPA, but if you open the designer, whenever a workflow is run (no matter whether it’s triggered locally or remotely), it will do a lot of work that will most likely cause memory leaks. So, during day-to-day operations, make sure you don’t have any designer open (the actual workflow).
Hmm, memory usage is tricky. First of all, .NET is a garbage-collected environment, meaning objects get released over time, and the VM needs “overhead”… so even if you just have a small spike for a short period of time, it can take a long time before that memory is released, if ever. The use of DataTables and loops is the most common way I see memory skyrocket, but when I talk about skyrocketing, I normally think 4-8 gigabytes of RAM and above. 1 gigabyte isn’t that bad in a .NET environment. If you know what you’re doing and you are fanatic about it, I have helped build robots that run thousands of workflows every hour and never go above 700 MB, so it can be done, but it requires lots of planning.
