This report summarizes the number of monitored nodes, the number of events that have been triggered in the last 24 hours, the number of affected nodes (nodes which have events related to them), percentage of events by criticality and average number of events per affected node.
The default "Triggered Alerts - Last 30 Days" report is the base of this report. You would need to prefix the NetPerfMon message in the alerts' trigger condition with the keywords 'Critical' or 'Warning' to get the criticality percentage in the report.
This report is provided "as is". You can optimize the query and modify the report as you want.
EDIT: Query modified to consider only node-specific alerts. Thanks Chris.
TODO: Put in a check to see if there are no events to avoid 'divide-by-zero- error.