Currently two knobs: samplingConfidenceInterval and samplingMarginOfError are used to control the CE sample size. This is confusing to everyone involved: engineers, testers, support team, and customers.

It also gives a false sense of confidence in the results of sampling CE. For example, the default values: "confidence interval = 95%" and "MoE = 5%" are based on certain statistical assumptions that are not always met in practice. 

This ticket should implement an optional possibility to define the exact number of samples to be drawn. If this parameter is set the Margin of error and the confidence interval should not be considered to determine the interval size.

This should be a global parameter which, in future, should be settable by PQS after PQS is able to handle any parameter.