Hi, I’m deploying a node.js application to Fly.io, we’re using OpenAI API to generate content for user, so the response is streaming in Server-Sent Events type, a normal response might take up to 10 seconds to complete.
I’d like to make sure if this workload works with auto_stop_machines=true, for example, will a request that is still streaming response to users be interrupted when the machine is stopped automatically?
Related configs in my fly.toml
kill_timeout = 10
[[http_service.checks]]
grace_period = "10s"
interval = "60s"
method = "GET"
timeout = "2s"
path = "/health"