What's changed?

Keep current-page selection available above and below Monitor Center's table before any monitor is selected. Both controls use the existing React selection controller, stay synchronized with row selection, and show a mixed state for partial selection. The bottom control shows the selected count and a clear action when applicable.

Selection continues to respect the current authoritative page, permissions, disappeared rows, and loading/error/empty states.

Closes #4425
Closes #4424

Validation

• Full frontend verification passed: toolchain, architecture, i18n, formatting, lint, type checking, unit tests, production build, bundle budget, dead-code analysis, and diff checks.
• Unit tests: 1,050 test files and 8,864 tests passed. The first full run exposed four ambiguous selectors in an existing page test after the second count/clear control was added; those assertions were scoped to the bulk action region, and the corrected four-test file passed on the final tree.
• Focused monitor list view suite: 30 tests passed.
• Real-backend browser checks: seven current-page monitors, selection in both directions, partial selection, clear, query changes, empty results, and refresh.
• Contributor confirmed local IDEA validation of the current React version: synchronized top/bottom controls, partial count, and clear selection.

Checklist

• I have read the contributing guide.
• Added regression tests for the changed selection controls.
• Documentation: no new public API or configuration contract.

Add or update API

No backend API or dependency changes.

AI assistance

AI assisted with implementation, tests, migration, and verification.

Human validation

The contributor reviewed and validated this React revision locally through IDEA before authorizing publication.

Risk notes

The real-backend instance contained seven monitors, so cross-page behavior was not exercised manually. The existing current-page selection semantics are preserved. Backend/release end-to-end checks were not run for this frontend-only change.