With the upcoming state sync feature (#4352) we will need to start a p2p stack with a limited set of reactors (e.g. only state sync and pex), and then enable the full set of reactors once the node is ready. While running a state sync, other peers will think that we are a "normal" peer, and attempt to use services that aren't available.

It would be nice if the node could announce which channels and service are available, so that peers can avoid sending messages to nodes that can't process them. This should probably have some sort of versioning scheme as well, to allow for reactor protocol changes.