Is there any summary or high level architecture drawing which would explain how network connection flows from outside network to a pod. What I would like to understand is will all traffic go through Rancher 2.0 server or not?
It depends on how the pod is being exposed; the various kinds of Service (ports on the workload in the UI), ingress rules, etc.
But none of them involve the server container, it is just for management and is not in the data-path of containers*. Clusters and the workloads in them continue running even if the server is down.
*: Unless you’re using the proxy in the management API to get to a service… that is explicitly through the server, but not a normal public way to get to something because it requires an auth token.
Thanks a lot. It’s like I expected, so i can say I have done something wrong with workload deployment since it’s not working.