Kong Hybrid deployment with Kuma Mesh

Hi Community

We are trying to deploy Kong in hybrid mode within K8s clusters, We have clusters enabled with Kuma mesh and some without Kuma mesh, we want to use these k8s clusters as Kong Data Plane.
Is there any reference architecture/best practices that we can follow to have uninterrupted flow of traffic between mesh enabled & disabled namespaces ?
In that case where should the Kong Dataplane pods should live ? within mesh or outside mesh ?