You can alsways submit a case. If you can reproduce the error with a smaller dataset, it is very useful to attach that. I would refer the older topic. If they know how to solve it, this is the fastest way on.

I’d start by checking a couple of things:

• Has anything changed in the data this workspace processes of late? Maybe size, complexity, type of geometry, number of vertices, that sort of thing?
• Disk space. FME uses quite a lot of disk space when dealing with larger datasets, so maybe it is running out of that. I’ve seen unexplainable error messages where the solution was to stop all FME instances, clear FME_TEMP, then try again.
• OS updates, changes, policies.

I’m saying this because the error message seems to point in the direction of space allocation, which runs out of room to do so. It could also have to do with a windows update, in a version that is a few years old, because the error message is coming from Boost.org, a C++ library for a bunch of things (including a container for geometry - so no surprise there, I guess). And from what I can see on Github, it seems to try and allocate more than is allowed/available/necessary, but that may also be caused by policies on windows (or your OS of choice) not allowing more than so much - which is why you wouldn’t see a spike when checking memory usage as a whole.

The size changes frequently and I do not think larger datasets have been an issue in the past, though I have been testing this and looking into this as that was my first thought as well. So far, nothing out of the ordinary. There should be plenty of disk space on this machine, but maybe something is wrong with the policy like you said. There was a windows update recently, perhaps that had something to do with it.
I am submitting a help ticket and also working with our IT team, I will let you know what they come up with. Thanks