I am working on a library (sorry, better documentation in English coming) with the following goal:
When having to process several / many work items, isolate the execution for each one against crashes by using distributed actors.
(Another goal is the ability to pause the execution for a work item, but this might not be the main point.)
The result should be a library that is flexible (i.e. applicable to similar use cases) and easy to use.
It would be interesting to know if there exist similar approaches (I did not find any) and if other people would be interested in such a library and maybe would like to give any suggestions or even join the efforts.
If your goal is only to isolate work so it can crash independently then indeed distributed actors is a nice way to achieve it but the cluster library may be a bit over an overkill for it -- it'll actually bind network resources etc.
It uses sub-processes and/or wasm to isolate workers which then could be allowed to crash in isolation etc. Please check it out and maybe we can collaborate on pushing that lib to "completion" from its prototype status?
PS: I would love to kickstart some form of "Distributed Actors Office Hours" soon so we could get together and push these things to completion, from just "proof of concepts" of which we have a lot... Maybe in the new year! FYI @jaleel
