So I'm torn between the timer being consistently wrong (it reports microseconds not milliseconds) or often wrong and occasionally correct. The part I like about being consistently wrong is that it makes it very easy to figure out what values are being reported and how to fix them. It also eliminates the chance of a metric being reported two ways, one call using Complete() the other CompleteAsMS(). Additionally, if at some point a metric was moved from micro Complete() to milliseconds CompleteAsMS() it would make analyzing historical data very difficult.
While annoying, I do believe that its best that we stay consistent in the timer metric we report.
So I'm torn between the
timerbeing consistently wrong (it reports microseconds not milliseconds) or often wrong and occasionally correct. The part I like about being consistently wrong is that it makes it very easy to figure out what values are being reported and how to fix them. It also eliminates the chance of a metric being reported two ways, one call usingComplete()the otherCompleteAsMS(). Additionally, if at some point a metric was moved from microComplete()to millisecondsCompleteAsMS()it would make analyzing historical data very difficult.While annoying, I do believe that its best that we stay consistent in the timer metric we report.