Another thing I was thinking was having different plans for each type of deployment/Repl.
One with a high Egrees limit (for Static Repls), one like Core with high storage, Autoscale compute units, & PostgreSQL, one mainly for bots with Reserved VMs, and maybe one with everything (like current Core but with Reserved VMs)
If you’re hosting a 24/7 website, autoscale will be MUCH more expensive than just having a reserved VM (and not necessarily better).
Here you can see the approximate price for the lowest tier autoscale deployment being used 24/7 for one month with Core. Very pricy when compared to a reserved VM (lowest tier is $6.40 a month without the additional $20 for Core)
I agree, autoscale is almost always the better option for sites. What I was saying was that if a site is receiving requests 24/7, it will be much more cost effective to use a Reserved VM. That’s a very uncommon scenario for the vast majority of sites though.