Describe logic of execution segments in distributed context

[yorugac]

Distributed execution inevitably adds complexity and sometimes there is a need in additional information about how the workload is split between runners. Sample questions:

• Reliably targeting some RPS, and in stages? #120
• Question: How does this project split rate between pods? #209

It'd be good to add a documentation page that highlights the key points about split of execution. Try to identify the pitfalls and appropriate recommendations.