Hi there,
I read the fly.io clustering guide and had some follow up questions:
- How does
fly scale count 2
work under the hood? - Is it possible to start a worker node from an existing main app node? Perhaps leveraging erlang :peer with the proper DNS node name.
- Would it be possible to start worker nodes on same host for vertical scaling AND different hosts for horizontal scaling?
For my use case I’m imagining an app that runs a single file exs script with its own Mix.installs. For dependencies to work the script needs its own isolated node and return output back to main app node. This is different yet related to app scaling / clustering since it could use libcluster.
Thank you and appreciate any help!