Would be great to have this in combination with a per-environment/service cost budget (instead of for the whole account), so the environment would get paused once it hits the budget limit.

I need something like this as I teach coding classes and the service only needs to be operating during the class, but removing the service is annoying, and sometimes I forget.

Serverless like GCP cloud run, if a web server or container hasn't received any requests/interaction then the it sleeps automatically, soon as a request comes in it wakes, preferably without much of a cold start.

Basically auto scaling but scales to zero without traffic, another example is AWS which has an equivalent called apprunner. They've monotized by charging when connecting to Github for auto deployments so they make something per month. Its something like $2 usd per container with auto deployment from github. Interesting model as the container can scale down but they make something anyway.

I'd love to be able to do this on an environment level, so I don't have to pay for my staging environment 24/7 but also don't have to reconfigure it every time I want to recreate it.

I would love to configure a web service, but not keep it running. Instead, I would rather log in to railway and press some button to start it back up. When I'm finished using it, I could pause it to prevent unneeded usage. The auto sleep feature would be nice, but I would be fine with either.

Not moving to In Progress, but we think we have an elegant answer to this that I think people would want and understand.

This will come in line with adding spend caps (above $20) and resource limits to manage spend with introducing app sleeping.

This feature would be the one to make me migrate some projects to Railway ! Ideally, with a programatic interface (e.g. schedule pause from monday to friday from midnight to 6 a.m.)

You can always manually remove the active deployment.