Open Martinsos opened 1 year ago
What does it even mean on Fly that 2 machines are used for server -> are they on at the same time? They have this thing where they shut down machines when they are not used -> so I guess only one of them is on normally? What if both are on? Is there some load balancing happening? Are they both used sometimes? In that case that might be an issue, if Wasp server is written in such a way that doesn't assume multiple server instances to be running (e.g. has some state).
Here's Fly's bit of docs on the topic https://fly.io/docs/reference/app-availability/#redundancy-by-default-on-first-deploy
We should probably expose the option to use --ha=disable
https://fly.io/docs/reference/app-availability/#turn-off-redundancy-on-deploy
I discussed this a bit more with @infomiho . What we will probably want to do is make it so by default 1 instance/machine of everything is created, and then inform user how they can bump it to 2 if they want. Once we get closer to 1.0 and invest proper effort into ensuring Wasp is horizontally scalable (docs, software support), then we can think more seriously about 2 machines being the default.
They create 2 machines for server, which is their recommended minimum, so you have availability while deploying new version of server and similar. But, I do wonder if this is what people expect -> it certainly surprised me quite a bit. Suggestions:
wasp
by default create only one server machine (but then again we should inform users about this choice).