Is it safe when using the Python Concurrent Futures to use ALL the vCores available in the machine?

My code currently uses the vCores available minus 1 as I want to leave one vCore “free” for the system, to avoid freezing the system. I am wonder if this approach is silly and if the concurrent futures manages somehow manages this.

def get_cpu_cores():
    """
    Get the number of CPU cores minus one.

    Returns:
        int: Optimal number of threads.
    """
    cpu_count = multiprocessing.cpu_count()
    return max(1, cpu_count - 1)