I see CPU spikes every minute which are caused by the statistic option of Openfire Enterprise.
In former times RRD did log to a local file (Current Users, Active Conversations, Packets Per Minute, ...) and now this information is stored for the whole cluster in the database. So every minute about 450 kB are updated in the database. This has little impact on my test installation but I wonder if this is good for larger installations or other databases.
Does one see similar CPU or network spikes?
Does one see a lot of archlogs using Oracle or similar things using another database which supports logging?