Machines REST API - Differences with fly.toml config


I’m still trying to get familiar with the machines API and it seems incredibly useful! I still have some questions regarding the API models when deploying/updating machines.

  1. Is there a way to set services.ports.tls_options when deploying machine apps with the REST API? I tried to set it along with the other properties in the services.ports section, but it seems to be ignored.

  2. The response from the create machines API has an init section that I couldn’t find any documentation for on the Machines API Reference docs. Could this be used to do something like the deploy.release_command from the fly.toml config?

  3. Will the processes-section eventually also be supported on the machines API, or would you always need seperate machines for stuff like background workers? I understand, this would be challenging with shutdown of one process on the machines.

I hope I didn’t miss any other topics covering these questions already.

Thanks in advance!


What TLS options would you like to set?

init is a way to over-ride the cmd/entrypoint I believe (instead of being a replacement for release commands). I don’t believe release commands are implemented yet.

The process section: You’re in luck, this was literally just posted: Multi process machines - try it out!

1 Like

Awesome, thanks for the quick reply!

Regarding the TLS options. For example: Restricting the proxy to only allow TLSv1.3.