Error message doesn't match getting started

[bmann]

We describe using the getting started repo which contains a settings.toml. If Homestar is not running when we visit the control panel, we get:

"Failed to connect to Homestar. Please check the WebSocket endpoint in your .env and ensure Homestar is running."

We should probably reference settings.toml here rather than .env

[avivash]

This seems potentially error-prone, since there are a few different ways to install Homestar. Maybe a good alternative is updating the getting started steps to ensure users set the correct value in their .env file?

[bmann]

There is no env file in our getting started instructions or workspace. Is this maybe for running control locally?

The install method doesn’t matter AFAICT — everyone runs a Homestar start that points at a settings.toml.

[avivash]

The .env is used when running the control panel both locally and in the cloud. It contains the CID for the uploaded workflow wasm and the ws address of the homestar node to connect to. also the fission server uri once we merge the auth branch.

Ryan and I have chatted before about having a way for users to update the address of the homestar node from within the control panel, so maybe that's something we need to look at adding soon. Otherwise I'm not sure of the best way for users to point the CP to the correct resources if they're running it from control.everywhere.computer 🤔

[avivash]

to simplify things we could also look at running the CP locally using a .toml file instead of a .env file. Though a .toml file may be foreign to a lot of Node users 🤔

[bmann]

Thanks for the background @avivash. Just logged it while it was fresh in mind, I think we can let this cook / bake for a while and see what feedback we get. And, "hosted Homestar" will be in the mix here too!