When one of my nodes has a problem, I would expect that this problem will be reflect up in a hierarchic map tree.
For example: when a switch in level 1 map has a CPU load of 80% and the warning threshold is 75%, then I expect to see the switch in warning state, the level 1 map symbol on the level 2 map also, and so on up to the top level map.
Can somebody tell me how I have to configure my maps?
It looks like a node's problem (e.g. CPU load) will be considered as a problem of a child object and will no be reported upwards.
So node status roll-up is focused on node availability. If on maps you want to view things like applications, IPSLA operations, volumes etc. you can add those elements to maps