I've got a Python application that I'm serving via IIS using ISAPI-WSGI. I've got it in its own application pool so that I can restart the Python interpreter whenever I want without disturbing other applications. I don't expect any more than 2 or 3 people to access it at any given time. Is there any reason for me to change the max number of worker processes to greater than the default of 1?
Also, what criteria should I consider when deciding how many worker processes to allocate? Is it just about concurrency or are there other factors?
Best Answer
One worker process is enough in my opinion. Consider that requests are queued, not rejected, when another request is being served and no more threads are available in the process (I'm assuming that the process is multi-threaded).