"Scaling up cluster" when going to a lighter service #1994
Description
I saw these logs when building a pipeline in sim4life.io
The two services, connected sequentially, are an osparc python runner (using 16 CPU 32 RAM) and a jsonifier (4 CPU 16 GB RAM). Therefore, I would expect the later to fit perfectly inside the machine the former just left free. Or is this made as always spin a new machine unless they have EXACTLY the same requirements?
In that case, might be useful to give more pricing options to the jsonifier so they all use the same machine - would that make things faster?