Is there a way to monitor the busyness of my dedicated server cpu cores and stop serving new connections passed a given cpu load threshold?
Put another way, what is the standard approach/technique to dynamically limit the maximum number of connections my machine can cope with?