learning, at a network device, a historic network traffic flow pattern related to traffic within a network, wherein learning the historic network traffic flow pattern related to traffic within the network comprises learning the historic network traffic flow pattern related to traffic within the network using a machine learning model;

based at least in part on the historic network traffic flow pattern, pre-defining, by the network device, two or more pre-defined modified rates of network control plane traffic for at least one class of traffic of a plurality of classes of traffic, wherein the two or more pre-defined modified rates of network control plane traffic are with respect to default rates of network control plane traffic for the at least one class of traffic of the plurality of classes of traffic;

receiving, at the network device within the network, network control plane traffic destined for a central processing unit (CPU) of the network device;

based at least in part on a current factor related to the historic network traffic flow pattern within the network, proactively selecting, by the network device, one of the two or more pre-defined modified rates of network control plane traffic for each class of traffic of the plurality of classes of traffic to provide a selected current rate of traffic for each class of traffic, wherein the current factor comprises one or more of a time of day, a day of week, a month, a date within a month, or a season;

implementing the selected current rates of network control plane traffic in the network;

determining, by the network device, the class of traffic for each packet of the network control plane traffic to provide a determined class of traffic;

based at least in part on the determined class of traffic of a particular packet and the selected current rate of traffic for the determined class of traffic, one of (i) routing the particular packet to the CPU of the network device or (ii) dropping the particular packet;

based at least in part on changes of traffic within the network, altering, by the network device, the historic network traffic flow pattern, wherein altering the historic network traffic flow pattern comprises altering the historic network traffic flow pattern using the machine learning model; and

based at least in part on altering the historic network traffic flow pattern, altering, by the network device, the two or more pre-defined modified rates of network control plane traffic for the at least one class of traffic of the plurality of classes of traffic.