An excerpt from the NNMi 9.20 Deployment Reference:
Blocking Incidents using the trapFilter.conf File Suppose the number of incidents flowing through your NNMi management server reaches a rate that causes NNMi to block newly arriving incidents. When this happens, NNMi generates a TrapStorm incident, indicating that incidents are blocked. NNMi might also generate a major health message indicating that the incident rate is high and incidents are being blocked. To remedy this, you might try to use the nnmtrapd.conf file to block incidents from entering NNMi in an attempt to reduce the incident traffic. However, if you use the nnmtrapd.conf file approach, NNMi still uses these incidents to calculate the trap rate and to write to the trap binary store. By using the nnmtrapd.conf file approach, you only stop incidents from being created or stored in the database. See the nnmtrapd.conf reference page, or the UNIX manpage for more information.
There is a better solution to this problem than using the nnmtrapd.conf file. NNMi provides a filtering mechanism that blocks incidents earlier in the NNMi event pipeline, preventing these incidents from being analyzed for trap rate calculations or from being stored in the NNMi trap binary store. By adding device IP addresses or OIDs to the trapFilter.conf file, you can block these high-volume incidents and avoid incident volume problems. See the trapFilter.conf reference page, or the UNIX manpage for more information.
If you have just started off with using NNMi 9.20, this link might help you in getting most common info you will need: