fix(kubernetes): use pod stats for keepalive

[MarcStdt]

Summary

• remove Kubernetes Hubble runtime/config/Helm integration
• use kubelet pod stats via nodes/proxy for procedure-level keepAliveTraffic
• report Kubernetes RX/TX/traffic fields from /ports
• fail open during stats warmup or unavailable pod stats

Validation

• go test ./...
• go test -count=1 -timeout 20m -tags="integration docker" -v ./test/integration/docker
• go test -count=1 -timeout 30m -tags="integration kubernetes" -v ./test/integration/kubernetes
• helm template chart check for nodes/proxy and no removed env

[druid-infra]

Error: This repo is not allowlisted for Atlantis.