An approach for determining compression intervals generally involves determining a compression interval based upon frequencies of event instances. According to the approach, event data is received that specifies instances of an event. The event data is analyzed and a compression interval is determined based upon frequencies of the instances of the event. The compression interval is selected to provide a balance between insufficient compression (under compression) and poor performance associated with over compression. Event compression may then be performed using the compression interval and report data may be generated that specifies the occurrence of the event.