docs(networking): add production risk and config guidance for NodeLocal DNSCache

[changluyi]

What / Why

Add production-readiness and operational guidance to the NodeLocal DNSCache doc. The current page only describes installation; it does not warn about the operational risks that come with changing every new Pod's DNS path on a node.

Changes

• Production usage risk warning at the top of Important Notes: default health-check port `8080` conflicts with common workloads, and a failed `node-cache` Pod breaks DNS for all Pods on that node without automatic failover to CoreDNS.
• Default Health Check Port: how to check `8080` usage (`ss -ltnp`), three remediation options, and a note not to hand-edit the generated ConfigMap/DaemonSet long-term.
• DNS Availability Risk: monitoring/alerting, upgrade maintenance window, emergency recovery (revert kubelet `cluster-dns` to the CoreDNS ClusterIP and recreate affected Pods).
• Optional: Configure Multiple DNS Servers: set CoreDNS ClusterIP as a secondary `cluster-dns` so new Pods get both nameservers, with glibc vs musl failover behavior notes and IPv4 test-cluster observations.

Notes

• No Chinese version of this file exists; only `docs/en/...` is updated.
• Out of scope: installation steps unchanged.