Describe your feature request
Autoscale deployments have spending limits so only so much credits or money can be spent.
What problem(s) would this feature solve?
It will stop people from being charged a lot of money because their deployment got a huge amount of traffic very quickly taking up a lot of computational units.
While the idea of implementing spending limits for autoscale deployments is beneficial to users, how would Replit maintain profitability if it stops charging when spending limits are reached due to increased traffic?
I personally would rather have Replit kill my deployment than charge me lots of money for using too much of a resource.