Convert SMTP uptime to days?

Having worked with several uptime-like datapoints recently, I think the issue could more easily be solved by giving datapoints a Unit of Measure. If the unit was known to be seconds, the graph could auto-scale when the number grows. For example, one datapoint i recently build gives the number of seconds the container has been up and running. It works really well, except when I’m dashboarding. Because the unit is seconds, the value gets pretty large pretty quickly. It would be nice if it auto scaled to hours, then days, then weeks, especially in big number widgets where there is only one number to worry about. Obviously, it’s a bit more complicated in a line graph were some would scale to minutes and others would scale to weeks, but that’s also doable. 

Of course it’s hard. If it was easy, I wouldn’t need LM to build it.  --Me in an actual conversation with an LM PM this week.