2024-01-06T00:18:22Z info: Deployment: c1c60d7a-93c0-4328-9632-ccb6781de116
2024-01-06T00:18:22Z info: Build: 9d3a1406-2fb5-4223-89ff-c369b18a8554
2024-01-06T00:18:22Z info: Connecting to builder
2024-01-06T00:18:24Z info: Builder connected
2024-01-06T00:18:24Z info: Installing packages
2024-01-06T00:18:24Z info: Skipping Build as no Build command set
2024-01-06T00:18:24Z info: Pushing pid1 binary layer...
2024-01-06T00:18:26Z info: Created pid1 binary layer
2024-01-06T00:18:27Z info: Skipping Node layer creation
2024-01-06T00:18:27Z info: Pushing Repl layer...
2024-01-06T00:18:27Z info: Retrieved cached nix layer
2024-01-06T00:18:27Z info: Created hosting layer
2024-01-06T00:18:30Z info: Created Repl layer
2024-01-06T00:18:30Z info: Pushed image manifest
2024-01-06T00:18:31Z info: Pushed soci index manifest
2024-01-06T00:18:35Z info: Creating Autoscale service
2024-01-06T00:19:48Z error: The deployment failed to initialize due to a configuration or code error. Check the logs tab for more information. For additional support, please post on https://ask.replit.com/c/help/13 and a staff member will take a look.
ok, and what does one do for bots that should be run if anybody is connected (or has been recently) but not run if nobody is using it, so that no processor time is wasted by repl.it (much less billed to anybody)
you know, the behavior of running a bot behind the existing repl behavior up till now, without Always Online or any other bloat?
i specifically do not want to put this in a reserved VM–i specifically want it to be spun down if it is not being used
what i want is best-effort hosting, and if hosting it would require more resources than repl.it can afford to give without paying more than i already am, it should just be throttled
hi, i already have a port being listened on, all of the usual works from back in the era when keepalive services were promoted, sans using a keepalive.
i literally just want the old behavior of best-effort hosting upon request that wont split the process or bill me for overages–i want it to just not use resources that would be billed to overage
as for not splitting the process i assume setting “maximum instances” to one does that?
i actually have multiple projects that i’ve had rely on the same behavior from repl.it–the common thread is that i use python to listen on a port (often websocket, and then returning a fixed http response on the root so repl.it considers the repl “up”) and communicating with some number of clients over said websocket
the bot and some other projects connect externally to eg discord, but some literally are just central hub servers for whoever happens to be currently connected, if anybody is–when i connect and get “hang on, the repl is loading” from my interface i know that nobody else has it open or just disconnected.
im concerned that if i have a spike in traffic, whereas historically the repl would have simply not kept up (which is what i want to happen) i will now be billed for the extra traffic as it handles it
is this the kind of elaboration you meant, or did you need other elaboration?