This is especially important with a service such as Heroku, where users will be changing these settings all the time and so there's a high probability of human error.
This is especially important with a service such as Heroku, where users will be changing these settings all the time and so there's a high probability of human error.
Akamai comes to mind as an example of a company that has always had similar issues. If I ever were to use them in the future, I would have to factor into their monthly costs the partial FTE required to manage their usage and costs. More importantly, if there are other CDNs that offer better cost management features, I would almost certainly use them before ever using Akamai again, unless there was no other technical option.
Perhaps you intentionally provisioned a large amount of headroom for a forthcoming expected/possible traffic spike.
In that example it would be annoying to get "are you sure you want to have this extra resource running?" emails.
(sure, you should be building the system to auto-provision but that would have required a better programmer and you just got the interns to build it, etc)