feat: refine cutover operations board and monitoring performance

Protect previous wave coverage, normalize batch identities, cache shared metric evidence, and batch diff inserts. Add live continuity KPIs, server paging and filters, scoped business selection, and preserve legacy monitor rule metadata.

Validation: 198 backend tests passed with 1 opt-in benchmark skipped; 14 frontend unit tests and 24 browser checks passed; TypeScript and production build passed.

tests/test_biz_migration_integrity.py

@@ -197,6 +197,84 @@ def test_port_mapping_does_not_remap_business_key_with_same_text():
     assert out["progress_ok"] == 1
 
 
+def test_unmapped_same_name_target_is_not_used_as_mapped_target_evidence():
+    out = evaluate(metric_id="interface_brief", key_fields=["port"], compare_fields=[],
+        old_baseline_rows=[{"port": "old"}], new_current_rows=[{"port": "old"}],
+        expect=parse_expect_set({"ports": ["old"]}))
+    row = next(r for r in out["rows"] if r["in_expect"])
+    assert row["verdict"] == "lost"
+    assert row["new"] == {}
+    assert row["new_key"] == ""
+    assert row["match_new_key"] == "new"
+
+
+def test_legacy_raw_interface_expect_uses_same_alias_normalization():
+    out = evaluate(metric_id="interface_brief", key_fields=["port"],
+        old_baseline_rows=[{"port": "GE-1"}], new_current_rows=[{"port": "GE-2"}],
+        compare_fields=[], port_map={"gei-1": "gei-2"},
+        expect=parse_expect_set({"ports": ["GE-1"]}),
+        iface_normalize_rules=[{"from": "GE", "to": "gei"}])
+    assert out["progress_ok"] == 1
+
+
+def test_baseline_picker_matches_normalized_keys_but_displays_originals(db):
+    ct = db.get(BizCompareTemplate, "ct")
+    ct.metrics_json = [{"metric_id": "arp", "sheet_id": "s", "key_fields": ["ip", "port"], "iface_fields": ["port"]}]
+    ct.iface_normalize_json = [{"from": "GE", "to": "gei"}]
+    db.get(BizStateMetricRow, "row-ob").data_json = {"ip": 0, "port": "GE-1"}
+    db.commit()
+    item = svc.list_baseline_expect_objects(db, "p")["sheets"][0]["items"][0]
+    assert item["keys"] == ["0", "gei-1"]
+    assert item["label"] == "0|GE-1"
+    assert parse_expect_set({"items": [{"metric_id": "arp", "key": item["keys"]}]})["arp"] == {"0\x1fgei-1"}
+
+
+def test_empty_composite_segments_are_preserved_and_false_is_not_blank():
+    assert parse_expect_set({"items": [{"metric_id": "x", "key": [0, "", False]}]})["x"] == {"0\x1f\x1ffalse"}
+
+
+def test_missing_status_field_cannot_match_negative_success_condition():
+    assert not _eval_field_op({"ip": "a"}, "state", "ne", "down")
+    assert not _eval_field_op({"ip": "a"}, "state", "not_in", ["down"])
+    assert _eval_field_op({"ip": "a"}, "state", "empty", None)
+
+
+def test_stable_outside_inventory_is_counted_without_repeated_evidence_cards():
+    row = {"ip": "a", "mac": "good", "port": "old"}
+    out = evaluate(expect={}, old_current_rows=[row], new_current_rows=[])
+    assert out["rows"] == []
+    assert out["steady_outside"] == 1
+    assert out["anomaly"] == 0
+
+
+def test_removed_previous_wave_sheet_blocks_full_coverage_acceptance(db):
+    db.add(BizMigrationBatch(id="prior", project_id="p", status="done",
+        created_at=utcnow_naive() - timedelta(days=1),
+        expect_set_json={"items": [{"sheet_id": "removed-sheet", "metric_id": "arp", "key": "old-customer"}]}))
+    db.commit()
+    finished = svc.finish_batch(db, "wave")
+    assert not finished["accept_summary"]["passed"]
+    assert finished["run"]["summary"]["uncovered_scope"] == ["removed-sheet"]
+
+
+def test_split_sheets_share_loaded_rows_and_command_metadata(db):
+    from unittest.mock import patch
+
+    ct = db.get(BizCompareTemplate, "ct")
+    sheet = ct.metrics_json[0]
+    ct.metrics_json = [sheet, {**sheet, "sheet_id": "s2"}]
+    for i in range(500):
+        for bid in ("ob", "nc"):
+            db.add(BizStateMetricRow(id=f"bulk-{bid}-{i}", batch_id=bid,
+                batch_command_id="cmd-" + bid, metric_id="arp", data_json={"ip": f"bulk-{i}", "mac": "good"}))
+    db.commit()
+    with patch.object(svc, "_load_metric_rows", wraps=svc._load_metric_rows) as load, patch.object(svc, "_command_brief", wraps=svc._command_brief) as command:
+        run = svc.run_evaluate(db, batch_id="wave")
+    assert load.call_count == 4  # once for each (batch, metric), not once per sheet
+    assert command.call_count == 2  # baseline + current metadata, independent of row count
+    assert svc.list_run_diffs(db, run["id"], limit=1)["total"] == 1002
+
+
 def test_wave_acceptance_rolls_back_whole_transaction_on_ticket_failure(db):
     from unittest.mock import patch