Hi Laf,
I guess there is an easy solution to this 
We just assume 100% uptime is the normal and should be a flat line instead of a constantly increasing graph. Then we just visualize the deviation from normal as a drop to zero. What I am thinking about is the uptime as an expression of 9’s per year or better say we just show the downtime.
See here: https://en.wikipedia.org/wiki/High_availability under Percentage calculation.
I guess there is a point if we should visualize:
- Downtime per year
- Downtime per month
- Downtime per week
- Downtime per day
but my preference would be “Downtime per year”
Every operator is proud of his “Nines” so maybe we can build an aggregate for the dashboard as well?