Record the ASG min-size and amount of time it is shutdown in logs/audit so we can calculate cost-savings.

[lyricnz]

When attempting to calculate the cost-savings of revolver (using external processes) over a given period:

• For vanilla EC2/RDS, we can calculate the average hourly cost from the billing data (for the hours it is up), then multiply that by the number of hours not-up to figure out the savings. (assuming it would have been 24/7)

• But for ASG/Clusters we need to know how many unit-hours were saved (which is min-size * number of hours), which can't be determined from the billing data alone.

[lyricnz]

We can mostly just ignore ASGs and count the EC2s that were stopped along with the ASG, and assume that's the number/type of resources that are solved over the shutdown period. At least until we terminate instances instead per #277