I converted 30d chrontab examples from "0 5 /30 " to "0 5 /1 " to ensure they run in months without 30days. (e.g. run on the 1st of the month rather than the 30th).
Across the board we've basically just set each metric frequency to equal the metric period. There are reasons to suspect this is may turn out to be problematic (thanks Willie) -- such as the often noted statement that one should sample at twice the rate of the highest frequency you want to capture (ref: https://en.wikipedia.org/wiki/Nyquist_rate ). For example, should we sample (e.g. set metricFrequency) to half the metricPeriod? I have not addressed this.
.... moved to metricChrontabExample
This is for issue #44
Some things to note: