I'm a *nix Engineer so probably someone more expert than me in the Win world can have his/her say on the matter anyhow keep in mind that spikes in memory and cpu usage are pretty normal in a production environment specially when on the same server you are running heave duty stuff like an antivirus and Lotus Domino (notes is the client part
), as long as this spikes are not hanging anything or creating problems/bottlenecks I would just ignore them.
I imagine you are using a monitoring software (Scom? Nagios? Opsview? something else?) and you should spend some time with the monitoring engineers to tweak it and adapt to your baseline server performance.
What most admins and engineers usually ignore is the fact that a monitoring software is pretty useless with default values, like MS SCOM, as every server, infrastruscutre and environment is different and simply you cannot have a "good for all" solution.
So take some time to analyze the situation and in case change the monitoring rules.