When I run a brut force ssh login to the server, my ip adress is banned but I can sil interact with the server.
crowdsec itself is in charge of the detection and emitting decision, but the bouncer(s) are in charge of applying the decision (see FAQ)
You should take a look at the bouncers on the hub to find where and how you want to apply the decision. If you don’t know which one to chose, netfilter is a good default to protect the machine itself, while the nginx one might more relevant if you’re doing web centric things !
Note: future versions of crowdsec’s wizard will directly suggest installation of relevant bouncers, but not done yet ^^