The disclosure provides an approach for rate limiting packets in a network. Embodiments include receiving, by a rate limiting engine running on a host machine, a network event related to a virtual computing instance running on the host machine, the network event comprising flow information about a network flow. Embodiments include receiving, by the rate limiting engine, context information corresponding to the network flow, wherein the context information comprises one or more of a user characteristic or an application characteristic. Embodiments include determining, by the rate limiting engine, a priority for the network flow by applying a rate limiting policy to the flow information and the context information. Embodiments include providing, by the rate limiting engine, the priority for the network flow to a multiplexer for use in rate limiting the network flow.