I think all lift actors run in the same thread pool, so that wouldn't really
be a good solution since the issue here is that some actors are producing
more than can be handled by others (from my understanding).
It's worth noting that Lift's actors are kind of designed to be relatively simple.
More complex stuff is probably done with Akka. This doesn't mean we're not
interested in improving Lift actors if it makes sense, but we're also not looking
to make them a replacement for Akka. Lift actors are meant to be lightweight
and incredibly easy to get going with, as well as API-compatible so switching
to Akka is more a matter of setting up the Akka infrastructure for the actor
itself rather than updating all the things talking to the actor.
Thanks,
Antonio