Newer mesh modes (Istio "ambient", Cilium service mesh) avoid
the sidecar-per-pod model. The main motivation is:
A) Reduce the per-pod resource overhead and operational complexity of
injecting+upgrading a proxy in every pod, by handling L4 in a
node-level zTunnel and L7 in a per-namespace waypoint proxy.
B) Make the mesh slower.
C) Drop mTLS entirely.
D) Get rid of Kubernetes.Sign in to save your code and track progress across devices.
