I have Autoscale deployment why does my app stop working when the browser closes

Question:
Why does my replit deployment stops working when I close my replit browser session? Do I need to select another type of deployment or do something different?

could you elaborate and give the code/repl for us to fiddle with?

Thanks. I have a chatbot deployment under rpelit code, autoscale deployment, the VM is deployed and running, but as soon as I close the browser it just stops working. Is it the deployment type or how can I ensure the application keeps working even if I dont have th browser running? https://replit.com/@jdomenzain/Lead-Gen-GPT-Chatbot

1 Like

could send the deployment link?

https://lead-gen-gpt-chatbot-jdomenzain.replit.app

Oh yeah, Autoscale shuts down if not in use (So if the page isn’t open), you should use a Reserved VM on Background Worker mode for this.

Alternatively, you could theoretically attach a pinger to the site, but that’s highly likely to be unstable and go down a lot.

1 Like

Do you mean Reserved VM Web servers?

I just tried it and it still shuts down after I close the browser.

No? Guessing by the repl name (chatbot) I defaulted to assuming it’s a discord bot, which doesn’t need the webserver part.

Does the deployment stop, or does the dev Repl stop running?

THe dev Repl stops running

That would be expected, as deploying doesn’t keep the dev end up. The deployed end should stay up, even if the Repl is off.

1 Like

Yeah I think that did it, I changed the dev link for the deployment one in Voiceflow and it solved it. Rookie move. Thanks for the pointer

Actually it did not haha, the fraking thing stopped working after some minutes…

Is it on autoscale or reserved VM? If it’s autoscale, then it shutting down would be expected.

1 Like

Autoscale Deployments don’t really “shut down”. They are always ready to receive requests. If the autoscale deployment receives traffic, it spins up very quickly and serves your content.

Reserved VMs on the other hand are always running.

Well I mean it would be expected to pause between requests, and the way it’s receiving them might not be waking it up properly.