Hey! I’m doing some mind-mapping of a hypothetical SaaS-tool and trying to figure out if it’d be good on render.
It will be running in node.js and have persistent WebSocket connections; with traffic spikes I’d to be able to scale up the number of WebSocket-servers.
The question I have is, how does scale down work, and when/how is it triggered
- How does it notify my server that it’s time for a graceful shutdown? When scaling down I’d have to notify connected clients to re-connect.
- How long is the grace period for shutting down?