"kubernetlab" is basically just running vrnetlab on Kubernetes. There really isn't a whole bunch of magic required, just tying together services, deployments, and vrnetlab containers in a kubernetes-y way.
Ideally this must should might get fleshed out to suck in your inventory
(Ansible, $other) and use that to determine the hosts to spin up as well as the
links between nodes and then do any or all of:
- generate a dot graph for the deployment to use as the reference point
- spit out the needed yaml for deployments, services, etc.
- actually push the deployments to a k8s API endpoint of your choosing
For now, basic examples are provided with nxosv. I can't distribute packaged nxosv containers, so fix-up the docker registry path in the manifests as needed until I get around to making helm-type things happen.
The network-policy.yaml manifest is optional if you need/want to isolate your kubernetlab namespace from the rest of the world. I've found this to be helpful in cases where I want to test out deployments/scenarios using device names from a real/live environment but obviously want to avoid accidentally reaching out from inside the kubernetlab environment to the "real world".