[Paul-ish]

The fact that the dashboards and alerts have a delay sounds like there might be difficult consistency stuff going on. Many nodes need to coordinate their usage and billing. It may be a difficult problem, but solving billing problems might not really motivate anyone at the company. It's not a "cool" problem for engineers and not profitable for product.

[phkahler]

>> The fact that the dashboards and alerts have a delay sounds like there might be difficult consistency stuff going on.

I think that's true. It's easier to measure usage and aggregate that data after the fact than to meter it in real time and stop at a limit. Those are very different things. What happens if you hit the cap while running multiple processes spread across a cloud?

One improvement might be to throttle things as the cap approaches but that doesnt really change the problem at all. Do that and have provider eat any overages should solve it from the user point of view.

[tpetry]

Theres an easy solution: You set a limit and everytime a service needs to spend some many it allocates a small portion of the budget and after some threshold it will put unused money back to the budget. The only downside is that your spending limit will be reached to optimisticly, but i prefer that to paying thousands more than i wanted to. Knowing the system works like maybe a lower and higher threshold for the budget could be set.