Resource Management
Lona is based on multi-threading (more information: Asynchronous Code). That means that every hook of a lona.LonaView like handle_request() or handle_input_event() block one thread each until your business logic finishes.
When planning resources for your application keep in mind that every view that runs can use up to three threads at a time (one for handle_request(), one for handle_input_event() and one for messaging between server and client).
Lona splits threading up in two pools: the view_runtime_pool and the worker_pool, configured through settings.MAX_RUNTIME_THREADS and settings.MAX_WORKER_THREADS.
LonaView.handle_request(), the main entry point for view business logic, runs in view_runtime_pool separated from all other Lona tasks because these are expected to run for a very long time potentially (up to days or weeks).
All other hooks and messaging runs in worker_pool. Business logic that runs there is expected to don't run that long, but with a much higher rate.
Example settings
These are example settings for an application that servers 50 concurrent users with one running view for each user. 50 views can run at a time. Because a view can use up to two worker threads the worker_pool should be at least twice as big as view_runtime_pool.
When running as a Lona script, static files like CSS, Javascript etc. get served by Lona itself. settings.MAX_STATIC_THREADS configures how many threads handle static file serving.
MAX_RUNTIME_THREADS = 50
MAX_WORKER_THREADS = 100
MAX_STATIC_THREADS = 20
Using Multiple Processes
Because of the Python Gil only one python thread can run at a time. Depending on your application and hardware it can make sense to start the same Lona project or script on multiple ports and use a load balancer like Apache2 to distribute the load to multiple Python processes.
All Apache configs in this article use Apaches load balancing feature to implement reverse proxying. To expand the balancing pool just add more than one member on port 8080.