diff --git a/docs/release-control/v6/internal/subsystems/alerts.md b/docs/release-control/v6/internal/subsystems/alerts.md
index 17ac19400..1549f436a 100644
--- a/docs/release-control/v6/internal/subsystems/alerts.md
+++ b/docs/release-control/v6/internal/subsystems/alerts.md
@@ -1390,6 +1390,16 @@ health checks, and storage offline lifecycle handling; future storage alert
 behavior should extend that resource checker owner while shared storage-health
 assessment helpers remain package-level until host and storage health paths are
 separated cleanly.
+Storage connectivity status is trimmed and case-normalised before evaluation.
+Empty or `unknown` status is absent connectivity evidence: it must not advance
+recovery or replace an existing incident's identity/start time. Capacity and
+pool-health evaluation remain independent and may still use valid observations.
+The existing non-failing treatment of available, online, active, inactive and
+disabled statuses remains unchanged; explicit connectivity-disable policy still
+clears the incident. `TestStorageUnknownConnectivityDoesNotRecover` and
+`TestStorageKnownConnectivityRecoveryCompatibility` in
+`internal/alerts/alerts_test.go` pin unknown-observation preservation, independent
+capacity activation, normalised offline activation and confirmed recovery.
 Proxmox node alert evaluation now lives in `internal/alerts/node.go`. That file
 owns node metric and temperature projection, node offline lifecycle handling,
 host-agent deduplication bookkeeping, and instance-scoped node display-name
diff --git a/internal/alerts/alerts_test.go b/internal/alerts/alerts_test.go
index 1718ece0c..8e1bc091e 100644
--- a/internal/alerts/alerts_test.go
+++ b/internal/alerts/alerts_test.go
@@ -21041,3 +21041,58 @@ func TestStorageEmptyCapacityRecovery(t *testing.T) {
 		})
 	}
 }
+
+func TestStorageUnknownConnectivityDoesNotRecover(t *testing.T) {
+	for _, status := range []string{"", "unknown", " UNKNOWN "} {
+		t.Run(status, func(t *testing.T) {
+			m := newTestManager(t)
+			disableTestTimeThresholds(m)
+			s := models.Storage{ID: "storage-observation", Name: "backups", Status: "unavailable"}
+			id := canonicalConnectivityStateID(s.ID)
+			for range 3 {
+				m.CheckStorage(s)
+			}
+			original := *testRequireActiveAlert(t, m, id)
+			s.Status = status
+			s.Total, s.Used, s.Free, s.Usage = 1000, 990, 10, 99
+			for range 5 {
+				m.CheckStorage(s)
+			}
+			active := testRequireActiveAlert(t, m, id)
+			if !active.StartTime.Equal(original.StartTime) || m.GetResolvedAlert(id) != nil {
+				t.Fatal("unknown connectivity changed the incident")
+			}
+			testRequireActiveAlert(t, m, canonicalMetricStateID(s.ID, "usage"))
+			s.Status = "available"
+			for range offlineRecoveryConfirmationsStorage {
+				m.CheckStorage(s)
+			}
+			if testHasActiveAlert(t, m, id) || m.GetResolvedAlert(id) == nil {
+				t.Fatal("confirmed available storage did not recover")
+			}
+		})
+	}
+}
+
+func TestStorageKnownConnectivityRecoveryCompatibility(t *testing.T) {
+	// Inactive shared storage and disabled storage are intentionally not
+	// connectivity failures; retain the existing treatment of these statuses.
+	for _, status := range []string{"available", "online", "active", "inactive", "disabled"} {
+		t.Run(status, func(t *testing.T) {
+			m := newTestManager(t)
+			s := models.Storage{ID: "storage-known", Name: "backups", Status: " OFFLINE "}
+			id := canonicalConnectivityStateID(s.ID)
+			for range 3 {
+				m.CheckStorage(s)
+			}
+			testRequireActiveAlert(t, m, id)
+			s.Status = status
+			for range offlineRecoveryConfirmationsStorage {
+				m.CheckStorage(s)
+			}
+			if testHasActiveAlert(t, m, id) || m.GetResolvedAlert(id) == nil {
+				t.Fatal("known non-failing status did not recover")
+			}
+		})
+	}
+}
diff --git a/internal/alerts/storage.go b/internal/alerts/storage.go
index 281d4f885..29ef3bf09 100644
--- a/internal/alerts/storage.go
+++ b/internal/alerts/storage.go
@@ -73,6 +73,7 @@ func (m *Manager) CheckStorageWithCapacityTrend(storage models.Storage, trend Ca
 
 	// Check if storage is truly offline/unavailable (not just inactive from other nodes)
 	// Note: In a cluster, local storage from other nodes shows as inactive which is normal
+	connectivityStatus := strings.ToLower(strings.TrimSpace(storage.Status))
 	if thresholds.DisableConnectivity {
 		m.mu.Lock()
 		for _, resourceID := range resourceIDs {
@@ -82,12 +83,14 @@ func (m *Manager) CheckStorageWithCapacityTrend(storage models.Storage, trend Ca
 		for _, resourceID := range resourceIDs {
 			m.clearAlert(canonicalConnectivityStateID(resourceID))
 		}
-	} else if storage.Status == "offline" || storage.Status == "unavailable" {
+	} else if connectivityStatus == "offline" || connectivityStatus == "unavailable" {
 		m.checkStorageOffline(storage)
-	} else {
+	} else if connectivityStatus != "" && connectivityStatus != "unknown" {
 		// Clear any existing offline alert if storage is back online
 		m.clearStorageOfflineAlert(storage)
 	}
+	// Missing connectivity evidence must not count as a healthy observation.
+	// Capacity and pool health below remain independently observable.
 
 	// Check usage if storage has valid data (even if not currently active on this node)
 	// In clusters, storage may show as inactive on nodes where it's not currently mounted