CPU Overload with Tomcat (Reference Data)

Average CPU load of 70% (during a 24 hour period) was observed under following circumstances:

Amazon Instance type:	m1.large
Processor Architecture:	64 bit
# Virtual CPU:	2
Amazon Compute Unit:	4
Memory:	7.5 GB
OS:	Linux

Total Incoming HTTP Requests:

2,414,000

HTTP request number is during a 24 hour period.

During the same period, average memory utilization was at 32%.

CPU load fluctuated with a high of 183% at 17:00 UTC; memory utilization was constant.

Peak load

During the last 2 hours culminating to the peak load of 183% at 17:00 UTC, about 168,000 HTTP requests were received per hour.

Conclusion

Incoming HTTP requests of 100,000 to 120,000 per hour is most optimal for the system, which will keep the load at 70% - 80%.

This post hopefully provides a reference for systems with similar configurations, even though, code executed per request is expected to vary from system to system.