diff --git a/VERSION b/VERSION
index c69d5ce..7babc03 100644
--- a/VERSION
+++ b/VERSION
@@ -1 +1 @@
-1.1.108
+1.1.109
diff --git a/cmd/edgeguard-api/main.go b/cmd/edgeguard-api/main.go
index b05a932..1608a5f 100644
--- a/cmd/edgeguard-api/main.go
+++ b/cmd/edgeguard-api/main.go
@@ -60,7 +60,7 @@ import (
 	usersvc "git.netcell-it.de/projekte/edgeguard-native/internal/services/users"
 )
 
-var version = "1.1.108"
+var version = "1.1.109"
 
 func main() {
 	addr := os.Getenv("EDGEGUARD_API_ADDR")
diff --git a/cmd/edgeguard-ctl/main.go b/cmd/edgeguard-ctl/main.go
index 0776139..77e64e9 100644
--- a/cmd/edgeguard-ctl/main.go
+++ b/cmd/edgeguard-ctl/main.go
@@ -11,7 +11,7 @@ import (
 	"git.netcell-it.de/projekte/edgeguard-native/internal/services/setup"
 )
 
-var version = "1.1.108"
+var version = "1.1.109"
 
 const usage = `edgeguard-ctl — EdgeGuard CLI
 
diff --git a/cmd/edgeguard-scheduler/main.go b/cmd/edgeguard-scheduler/main.go
index 58ee47e..d1a1391 100644
--- a/cmd/edgeguard-scheduler/main.go
+++ b/cmd/edgeguard-scheduler/main.go
@@ -16,6 +16,7 @@ import (
 	"log/slog"
 	"net"
 	"os"
+	"os/exec"
 	"regexp"
 	"strconv"
 	"strings"
@@ -40,7 +41,7 @@ import (
 	"git.netcell-it.de/projekte/edgeguard-native/internal/services/tlscerts"
 )
 
-var version = "1.1.108"
+var version = "1.1.109"
 
 const (
 	// renewTickInterval — how often we re-evaluate expiring certs.
@@ -124,6 +125,14 @@ const (
 	conntrackCheckInterval = 2 * time.Minute
 	conntrackWarnPct       = 80.0
 	conntrackCriticalPct   = 90.0
+
+	// ntpSyncCheckInterval — alle 10 Minuten chronyc tracking aufrufen.
+	// Keine Sync bedeutet: Uhr driftet → TLS-Cert-Prüfung schlägt fehl
+	// wenn die Abweichung > Toleranz des Gegenstücks (i.d.R. ±1 min),
+	// JWT-Ablauf inkonsistent, Cluster-Split-Brain möglich. Dedupe 1h
+	// damit ein kurzer Upstream-Ausfall (Reboot, DHCP-Pause) keinen
+	// Alert-Regen produziert.
+	ntpSyncCheckInterval = 10 * time.Minute
 )
 
 func main() {
@@ -221,6 +230,12 @@ func main() {
 	defer conntrackTick.Stop()
 	runConntrackCheck(ctx, alertSvc, alertDedupe)
 
+	ntpSyncTick := time.NewTicker(ntpSyncCheckInterval)
+	defer ntpSyncTick.Stop()
+	// Kein Initial-Check bei Boot: chrony braucht nach dem Start
+	// einige Sekunden bis zur ersten Synchronisation — ein
+	// sofortiger Check würde immer feuern.
+
 	for {
 		select {
 		case <-renewTick.C:
@@ -250,6 +265,8 @@ func main() {
 			runMemoryCheck(ctx, alertSvc, alertDedupe)
 		case <-conntrackTick.C:
 			runConntrackCheck(ctx, alertSvc, alertDedupe)
+		case <-ntpSyncTick.C:
+			runNTPSyncCheck(ctx, alertSvc, alertDedupe)
 		}
 	}
 }
@@ -474,6 +491,78 @@ func runConntrackCheck(ctx context.Context, a *alerts.Service, d *dedupe) {
 	}
 }
 
+// runNTPSyncCheck ruft chronyc tracking auf und feuert einen Alert wenn
+// chrony keine synchronisierte Zeitquelle hat (Stratum 0 oder ≥ 16).
+// Zeitdrift > ~1 Minute führt zu TLS-Handshake-Fehlern, JWT-Ablauf-
+// Inkonsistenzen und möglichen Cluster-Problemen. Dedupe 1h.
+func runNTPSyncCheck(ctx context.Context, a *alerts.Service, d *dedupe) {
+	if a == nil || d == nil {
+		return
+	}
+	out, err := exec.Command("chronyc", "tracking").Output()
+	if err != nil {
+		// chrony nicht installiert oder nicht gestartet — kein Alert,
+		// weil wir nicht wissen ob chrony hier überhaupt erwartet wird.
+		return
+	}
+	synced, stratum, ref := parseChronyTrackingForAlert(string(out))
+	if synced {
+		return
+	}
+	const key = "ntp.unsync"
+	if !d.shouldFire(key) {
+		return
+	}
+	refStr := ref
+	if refStr == "" {
+		refStr = "(keine Referenz)"
+	}
+	title := fmt.Sprintf("NTP nicht synchronisiert (Stratum %d)", stratum)
+	desc := fmt.Sprintf(
+		"chrony hat keine synchronisierte Zeitquelle.\n"+
+			"Referenz: %s  Stratum: %d\n\n"+
+			"Mögliche Ursachen:\n"+
+			"  • Upstream-NTP-Server nicht erreichbar (UDP/123 blockiert?)\n"+
+			"  • Pool-DNS-Einträge lösen nicht auf\n"+
+			"  • chrony läuft, braucht aber noch Zeit nach Boot (warten)\n\n"+
+			"Prüfen: chronyc sources -v  —  chronyc tracking",
+		refStr, stratum)
+	if _, err := a.Fire(ctx, "ntp.unsync", alerts.SeverityWarning, title, desc); err != nil {
+		slog.Warn("scheduler: ntp-sync-check alert fire failed", "error", err)
+	}
+}
+
+// parseChronyTrackingForAlert ist eine schlanke Variante des NTP-Handler-
+// Parsers: liefert nur synced/stratum/reference ohne die vollen Felder.
+func parseChronyTrackingForAlert(out string) (synced bool, stratum int, reference string) {
+	for _, line := range strings.Split(out, "\n") {
+		line = strings.TrimSpace(line)
+		key, val, ok := strings.Cut(line, ":")
+		if !ok {
+			continue
+		}
+		key = strings.TrimSpace(key)
+		val = strings.TrimSpace(val)
+		switch key {
+		case "Reference ID":
+			if i := strings.Index(val, "("); i >= 0 {
+				reference = strings.Trim(val[i:], "()")
+			}
+			if val != "00000000 ()" {
+				synced = true
+			}
+		case "Stratum":
+			fmt.Sscanf(val, "%d", &stratum)
+			if stratum > 0 && stratum < 16 {
+				synced = true
+			} else if stratum == 0 || stratum >= 16 {
+				synced = false
+			}
+		}
+	}
+	return
+}
+
 var egBackendRE = regexp.MustCompile(`^eg_backend_(\d+)$`)
 
 // runBackendDownCheck liest HAProxy-Stats via Admin-Socket und feuert
diff --git a/management-ui/src/i18n/locales/de/common.json b/management-ui/src/i18n/locales/de/common.json
index 1ef7c76..0481374 100644
--- a/management-ui/src/i18n/locales/de/common.json
+++ b/management-ui/src/i18n/locales/de/common.json
@@ -1119,7 +1119,7 @@
     "title": "Health-Alarme",
     "intro": "Notification-Channels für kritische Events. Webhook (Slack/Discord/Teams/Generic-HTTP) oder Email (SMTP). Triggers: cert.expiring (<14 d), cert.renew_failed, backup.failed, license.invalid.",
     "scopeTitle": "Was triggert Alarme?",
-    "scopeDesc": "cert.expiring — TLS-Zertifikat <14 Tage Restzeit (dedupe 12h). cert.renew_failed — ACME-Renewer hat Fails. backup.failed — Scheduled Backup konnte nicht erstellt werden. license.invalid — License-Server liefert valid=false. backend.down — alle Server eines Backend-Pools sind DOWN (2-Min-Check, dedupe 12h). disk.full — Root-Filesystem ≥80% Warnung, ≥90% Critical (stündlich, dedupe 12h). mem.high — RAM-Auslastung ≥85% Warnung, ≥95% Critical (5-Min-Check, dedupe 1h). conntrack.high — Conntrack-Tabelle ≥80% Warnung, ≥90% Critical; bei 100% werden alle neuen Verbindungen lautlos verworfen (2-Min-Check, dedupe 1h).",
+    "scopeDesc": "cert.expiring — TLS-Zertifikat <14 Tage Restzeit (dedupe 12h). cert.renew_failed — ACME-Renewer hat Fails. backup.failed — Scheduled Backup konnte nicht erstellt werden. license.invalid — License-Server liefert valid=false. backend.down — alle Server eines Backend-Pools sind DOWN (2-Min-Check, dedupe 12h). disk.full — Root-Filesystem ≥80% Warnung, ≥90% Critical (stündlich, dedupe 12h). mem.high — RAM-Auslastung ≥85% Warnung, ≥95% Critical (5-Min-Check, dedupe 1h). conntrack.high — Conntrack-Tabelle ≥80% Warnung, ≥90% Critical; bei 100% werden alle neuen Verbindungen lautlos verworfen (2-Min-Check, dedupe 1h). ntp.unsync — chrony hat keine synchronisierte Zeitquelle; Drift führt zu TLS- und JWT-Fehlern (10-Min-Check, dedupe 1h).",
     "tabs": { "channels": "Channels", "events": "History" },
     "add": "Channel hinzufügen",
     "addTitle": "Notification-Channel anlegen",
diff --git a/management-ui/src/i18n/locales/en/common.json b/management-ui/src/i18n/locales/en/common.json
index 8192eec..11f899e 100644
--- a/management-ui/src/i18n/locales/en/common.json
+++ b/management-ui/src/i18n/locales/en/common.json
@@ -1119,7 +1119,7 @@
     "title": "Health alerts",
     "intro": "Notification channels for critical events. Webhook (Slack/Discord/Teams/generic-HTTP) or email (SMTP). Triggers: cert.expiring (<14 d), cert.renew_failed, backup.failed, license.invalid.",
     "scopeTitle": "What triggers alerts?",
-    "scopeDesc": "cert.expiring — TLS cert <14 days remaining (12 h dedupe). cert.renew_failed — ACME renewer cycle had failures. backup.failed — scheduled backup couldn't run. license.invalid — License server returns valid=false. backend.down — all servers in a backend pool are DOWN (2 min check, 12 h dedupe). disk.full — root filesystem ≥80% warning, ≥90% critical (hourly check, 12 h dedupe). mem.high — RAM usage ≥85% warning, ≥95% critical (5 min check, 1 h dedupe). conntrack.high — conntrack table ≥80% warning, ≥90% critical; at 100% all new connections are silently dropped (2 min check, 1 h dedupe).",
+    "scopeDesc": "cert.expiring — TLS cert <14 days remaining (12 h dedupe). cert.renew_failed — ACME renewer cycle had failures. backup.failed — scheduled backup couldn't run. license.invalid — License server returns valid=false. backend.down — all servers in a backend pool are DOWN (2 min check, 12 h dedupe). disk.full — root filesystem ≥80% warning, ≥90% critical (hourly check, 12 h dedupe). mem.high — RAM usage ≥85% warning, ≥95% critical (5 min check, 1 h dedupe). conntrack.high — conntrack table ≥80% warning, ≥90% critical; at 100% all new connections are silently dropped (2 min check, 1 h dedupe). ntp.unsync — chrony has no synchronized time source; clock drift causes TLS and JWT failures (10 min check, 1 h dedupe).",
     "tabs": { "channels": "Channels", "events": "History" },
     "add": "Add channel",
     "addTitle": "Add notification channel",