Earlier this week, AWS indicated our CSR 1000v instance needed to be rebooted for some maintenance. Since the reboot event, the AWS monitored CPU utilization went from about 11% to 65%. The CPU utilization reported by the router itself is about 5%. Neither the router software nor the configuration was changed, so I don't know why the sudden change in the monitored CPU usage.
The amount of data being serviced by this router is quite small: ~100KiB/s so it shouldn't tax the router.