Improve topology debugging UX and release tag consistency

[kaskol10]

Summary

• extend topology UX with compact/full pod list toggle, hide-daemonset toggle, node drill-in, daemonset-only mode, node-local search, and per-node persisted drill-in filters
• add CPU and memory request values to drill-in pod rows for faster debugging context
• fix local docker-compose defaults by mapping frontend to 3000:80 and mounting host kubeconfig into API container
• update release automation to bump both backend and frontend Helm image tags in values.yaml whenever chart version is bumped

Test plan

• cd frontend && npm run lint
• cd frontend && npm run build
• manual topology UI sanity checks (overview toggles, drill-down controls, filter persistence)
• validate compose config changes (docker-compose.yml) for local API/UI wiring

Notes

• Helm release workflow now updates charts/karpenter-optimizer/values.yaml tag fields so backend/frontend image tags stay aligned with chart version.