Looking at the architecture on https://docs.k3s.io/architecture#single-server-setup-with-an-embedded-db to try and understand really where the load balancer sits and operates from. My understanding is that

servers

actually run Traefik and this acts as the load balancer. I am planning to use a single server so I would like to understand if i actually need to spin my own load balancer solution, how and if not if indeed the loadbalancer is actually withing the Server Node

this one?

oh actually you answered

but the LoadBalancer is accessible on all the nodes

lets say my amazing app ... and run it on this k3s cluster

so just point it at 80/443 on all the nodes

to node(s) or to server(s)

great, more complex that what i used to use but these days it looks like i cant do it without Kubernetes

ServiceLB is actually pretty simple. Without it you would need to muck about with Kube-vip or metallb or something

and each app deployment can deploy its LB right ?

You can only use a port on a single service. If you try to use the same port for another service it will hang in pending until the other thing using the port goes away

ok so i indeed need an ingress

you have an ingress controller in Traefik. you need to create ingress resources to tell it how to forward requests to your services.

i remember ingress indeed. I had not clue though of this difference between Service LB and Ingress

Services are layer 3, Ingresses are layer 7.

i think i will need to spend sometime around these pages once i fix this TLS issue

oh yeah, 1.x is basically docker on steroids right? thats before my time.

Lol yes it is but you dont need to mess around more inception and other contexts that k8s brings