I have two deployments (A and B), each one exposing ClusterIP Service. Before deploying Istio, I was able to communicate from pod A to any of B pods via its Service (e.g. http://B.default.svc.cluster.local/dosomecrazystuff)
After deploying Istio (1.0.5), I getting "http://B.default.svc.cluster.local refusing connection" when calling it from pod in deployment A.
What is default routing policy in Istio? I don't need some cleaver load balancing or version based routing, just straightforward communication from A to B (the same way as I would do that without Istio).
What the absolute minimal required configuration to make it work?
Well, it seems like some local issue I having on my MicroK8s deployment. On EKS and another MicroK8s I able to communicate as desired without anything special.
So, the answer is: no special configuration required to make it work, it supposed to be able to communicate just as is.