The maintenance flag affects a node's willingness to accept a new job, but does not turn off job dispatch. Putting the admin node into maintenance would pause dispatch (really creation) of *new* jobs, but current jobs (ie, already created) would still be dispatched until they're all exhausted. Ingest would still work, however no new workflows would show up. I would bet that the ingests would not even show up in the admin ui, although I have not tested that. If you want to pause ingests, turn off your load balancer/proxy in front of admin, or add an exception to the redirect for all paths prefixed with /ingest. All well behaved CAs should store things locally until you're restored the default configuration. Misbehaving CAs.... well, let's just say you shouldn't test this with prod until you're sure about your CAs :)
G