I don't understand why there isn't at least a setting that says "turn everything off if I hit $x."
Then just given people a certain grace period to reactivate or get their data out before it's removed.
It wouldn't fix production deployments where you want alarms, not a shutdown, when you hit spending caps, but it would help people on the dev stage to avoid issues like this.
So let’s say they “turn everything off”. Does that include deleting all of your objects in S3? Deleting your database? Deleting your attached disks (AMIs)? Deleting your DNS entries?
>Does that include deleting all of your objects in S3? Deleting your database? Deleting your attached disks (AMIs)? Deleting your DNS entries?
People bring this every time. You give a grace period and then yes - assuming the user opted into a "turn everything off".
Its not some impossible engineering feat & clouds are full of "out of money data gone" stuff already. e.g.
>If a paid subscription ends or is terminated, Microsoft retains customer data stored in Microsoft 365 in a limited-function account for 90 days to enable the subscriber to extract the data.
If it a personal site I run for kicks and expect to cost less than $100/year, but is suddenly running into thousands? Yes please, delete it all. It's the only way I could sleep at night.
Sure. I'd enable that on dev / testing accounts without hesitation. I don't know why so many people pretend like everyone will be forced into getting resources deleted if there's an option of a hard limit. You can have multiple accounts. They even recommend it.
It makes way more sense for me to build my stuff using a dev/testing account. After that I'll have a good enough understanding of the resources I'm using that it's practical for me to configure more complicated cost controls using a production account.
That sounds like Amazon taking one of their problems and pushing it onto the users. If the root limitation is that Amazon's billing process is so poor that it can't interact with their other processes, then that should be Amazon's problem to fix. Until and unless Amazon fixes their own problem, Amazon should be eating the cost resulting from Amazon allocating server usage beyond a user-specified maximum billing.
Then just given people a certain grace period to reactivate or get their data out before it's removed.
It wouldn't fix production deployments where you want alarms, not a shutdown, when you hit spending caps, but it would help people on the dev stage to avoid issues like this.