Summary

Refactor the functional collector benchmark tool to run over a certain time duration in lieu of processing a set number of message.  The motivation is  a fixed number of messages may not realistically expose  how the collector consumes CPU, metrics and functions under load

Acceptance Criteria

• functional-benchmarker that is runs for a fixed duration
• Generates statistics including: mem, cpu, latency
• Capability to specify the config to be tested instead of relying on the CL Generator