There is no option for that by configuration only. You will need catch on this host such behaviour. On DT side you will need to create custom network device (for example using host name of this unmonitored one). Then you should pass those occurencies of calls as custom metric into dynatrace. On such metrics you can define custom thresholds. Then problem will be rised.
I fee your pain on this one. It's a hard one to find an easy answer to for both the customer and Dynatrace. I am still trying to figure out the best way to deal with unmonitored hosts. I might not be stating this exactly right, but from what I can tell Dynatrace treats "Unmonitored hosts" as one big service. The problem is unmonitored hosts can be connected to numerous applications, services, hosts, etc. from multiple teams who own different things. The 2 things we are trying to figure out are:
Actually I'm always splittiing unmonitored hosts into custom network device (now you can do it via UI). It is easier to tag it and assign to right management zone for example. But for the purpose @Bastien T. is talking about I have no clue... Maybe log analytics, writing such incidents into logs and creating alerts based on them?