This code snippet shows how to wrap a concurrent.futures.Executor class to provide a limited queue size.

max_queue_size_pool.py

from threading import BoundedSemaphore
from concurrent.futures import ProcessPoolExecutor


class MaxQueuePool:
    """This Class wraps a concurrent.futures.Executor
    limiting the size of its task queue.

    If `max_queue_size` tasks are submitted, the next call to submit will block
    until a previously submitted one is completed.

    """
    def __init__(self, executor, max_queue_size, max_workers=None):
        self.pool = executor(max_workers=max_workers)
        self.pool_queue = BoundedSemaphore(max_queue_size)

    def submit(self, function, *args, **kwargs):
        """Submits a new task to the pool, blocks if Pool queue is full."""
        self.pool_queue.acquire()

        future = self.pool.submit(function, *args, **kwargs)
        future.add_done_callback(self.pool_queue_callback)

        return future

    def pool_queue_callback(self, _):
        """Called once task is done, releases one queue slot."""
        self.pool_queue.release()


if __name__ == '__main__':
    pool = MaxQueuePool(ProcessPoolExecutor, 8)
    f = pool.submit(print, "Hello World!")
    f.result()

It's better to use from multiprocessing import BoundedSemaphore, not from threading, I guess. Because you're using it with ProcessPoolExecutor, not ThreadPoolExecutor.
Correct me if i'm wrong, please

multiprocessing semaphores work across processes but they are way heavier than threading ones as they involve OS IPC routines.

In the above case, we want to block the thread publishing data into the pool. Therefore, the threading semaphore is enough.

@noxdafox great, thank you!

@noxdafox could you explain some point, pls? I need to do some work upon the result of each future. What is the best way for doing this: to add extra callable to future.add_done_callback in the custom submit method or to use smth like as_completed() in the main pipeline? Is it possible to make it non blocking the queue filling?

Sorry but this is not a place for such discussions.

Please use stackoverflow for that. Create a question stating what you are trying to achieve, what you have been doing so far and where you are having troubles/being stuck. Please tag it as multiprocessing, I usually am pretty active there.