But if I do this I can’t scale horizontally with node replicas as I’m using redis to maintain the user state - thus rendering that capability a little redundant?
For horizontal scaling you might consider running Redis next to Traefik on the single entry server that proxies all requests to the working nodes.
We hope to add docker swarm support to the resource manager soon.
Your suggestion is a little beyond me now - which is why I was trying to use Wappler functionality of remote replicas plus a remote redid. Probably it should be solved so others can use this if they need to scale rather than just not using it? I would assume most users if they use replicas would need this to maintain user states.
I’ll learn about docker swarm - do you have an approx eta do I can realistically think of this as an alternative?
How did you plan to scale horizontally?
I feel like I’m about to go back and do some learning
The plan to scale horizontally is via node replicas and a larger, matched server using remote redis to manage user states regardless of which replica they are on.
I’ll be doing load testing on specific public and secure api’s that will get hit with the most traffic to gauge scaling needs initially and keep retesting based on growth.
You can run Redis globally as explained in:
Ok, thanks George. I did try this setup at the beginning which didn’t seem to work but I’ll give it another run and report back.