When using GKE, I found that a all the nodes in the Kubernetes cluster must be in the same network and the same subnet. So, I wanted to understand the correct way to design networking.
I have two services A and B and they have no relation between them. My plan was to use a single cluster in a single region and have two nodes for each of the services A and B in different subnets in the same network.
However, it seems like that can't be done. The other way to partition a cluster is using namespaces, however I am already using partitioning development environment using namespaces.
I read about cluster federation https://kubernetes.io/docs/concepts/cluster-administration/federation/, however it my services are small and I don't need them in multiple clusters and in sync.
What is the correct way to setup netowrking for these services? Should I just use the same network and subnet for all the 4 nodes to serve the two services A and B?
You can restrict the incoming (or outgoing) traffic making use of labels and networking policies.
In this way the pods would be able to receive the traffic merely if it has been generated by a pod belonging to the same application or with any logic you want to implement.
You can follow this step to step tutorial that guides you thorough the implementation of a POC.
kubectl run hello-web --labels app=hello \
--image=gcr.io/google-samples/hello-app:1.0 --port 8080 --exposeExample of Network policy
kind: NetworkPolicy
apiVersion: networking.k8s.io/v1
metadata:
name: hello-allow-from-foo
spec:
policyTypes:
- Ingress
podSelector:
matchLabels:
app: hello
ingress:
- from:
- podSelector:
matchLabels:
app: foo