How does one get classic hosting behavior from deployments:tm:

I have multiple old projects of mine (mostly python) that (up until the forced switch to deployments) have associated permanent URLs which when accessed result in some sort of “central hub” software running (eventually)–the classic hosting behavior.
I specifically have not wanted the behavior of Always Online for these projects, so I have not used that for these projects and am not looking to use “dedicated VM” to emulate the behavior of Always Online.
I am looking specifically to recreate the behavior of a single central process being run when triggered by request, without persistent filesystem (I do need a handful of files to be accessible, they are however not programmatically edited), and that stops running if not used for awhile; I specifically rely on detecting the “this repl has not been spun up yet” behavior to detect when the current client is “the first client”

I have no one repl link to link for this because it is a slew of old links i shall have to find because the links i have been using have just been broken by the change.

however I can provide other contextual information on how the majority of them work if I recall correctly: they tend to be python programs utilizing a small handful of python source files directly imported, listening on a single port, sometimes (but not always) connecting to a fixed external location for a small amount of data storage (hence not needing persistent filesystem) as well as to eg discord both as a bot and to trigger webhooks to let me know when something goes wrong.

I am under the impression that Autoscale Deployments:tm:, while free for an extent on Hacker/Core, will just start billing you if there’s a spike in traffic. how can i specify that if traffic would be such that it would create overages, to just throttle it instead?

Hey @eaz, the best way to get the result of the classic hosting behaviour is autoscale. Sadly there is no way of getting legacy hosting back. Have a Good one!