Initial commit: server-monitor v1 (Python)

Kompletter Neustart als Python-Projekt.
Portiert von zfs-monitor Bash v3, saubere Architektur:
- config.py: TOML-Config mit dataclasses
- state.py: zentrales State-Management (JSON statt Einzeldateien)
- notify.py: Telegram + Persistent Alerting als wiederverwendbare Klasse
- checks/zfs.py: ZFS Pool-Health, I/O-Fehler, Kapazität
- checks/smart.py: SATA + NVMe SMART-Monitoring
- checks/nut.py: NUT/USV-Monitoring (APC Back-UPS RS 1500G)

notify.py

@@ -0,0 +1,99 @@
+from __future__ import annotations
+
+import json
+import logging
+import time
+import urllib.parse
+import urllib.request
+from datetime import datetime
+
+from config import Config
+from state import State
+
+log = logging.getLogger(__name__)
+
+ICONS = {"CRITICAL": "🔴", "WARNING": "🟡", "INFO": "🟢"}
+
+
+class Alerter:
+    def __init__(self, config: Config, state: State) -> None:
+        self.config = config
+        self.state = state
+
+    # ── Persistent Alerting ───────────────────────────────────
+    # Gleiche Logik wie im Bash-Script:
+    #   - Neuer Fehler       → sofort melden, runs = 0
+    #   - Fehler bleibt      → runs++, melden wenn runs >= repeat_runs
+    #   - Fehler weg         → Entwarnung, State löschen
+
+    def should_alert(self, key: str, level: str) -> bool:
+        repeat = (
+            self.config.alert.critical_repeat_runs
+            if level == "CRITICAL"
+            else self.config.alert.warning_repeat_runs
+        )
+        active = self.state.get(f"{key}_active")
+        runs: int = self.state.get(f"{key}_runs", 0)
+
+        if active != "1":
+            self.state.set(f"{key}_active", "1")
+            self.state.set(f"{key}_runs", 0)
+            return True
+
+        runs += 1
+        self.state.set(f"{key}_runs", runs)
+
+        if repeat == 0:
+            return False
+        if runs >= repeat:
+            self.state.set(f"{key}_runs", 0)
+            return True
+        return False
+
+    def clear_alert(self, key: str) -> bool:
+        """Clear state and return True if the alert was previously active."""
+        was_active = self.state.get(f"{key}_active") == "1"
+        self.state.delete(f"{key}_active", f"{key}_runs")
+        return was_active
+
+    # ── Sending ───────────────────────────────────────────────
+
+    def send_alert(self, level: str, title: str, body: str, key: str = "") -> None:
+        icon = ICONS.get(level, "⚪")
+        suffix = ""
+        if key:
+            runs: int = self.state.get(f"{key}_runs", 0)
+            if runs > 0:
+                suffix = f" <i>(Erinnerung #{runs})</i>"
+
+        ts = datetime.now().strftime("%d.%m.%Y %H:%M:%S")
+        msg = (
+            f"{icon} <b>[{level}] {self.config.hostname}</b>{suffix}\n"
+            f"<b>{title}</b>\n\n"
+            f"{body}\n\n"
+            f"🕐 {ts}"
+        )
+        log.info("ALERT %s: %s", level, title)
+        self._send_telegram(msg)
+
+    def _send_telegram(self, text: str) -> bool:
+        url = f"https://api.telegram.org/bot{self.config.telegram.token}/sendMessage"
+        data = urllib.parse.urlencode({
+            "chat_id": self.config.telegram.chat_id,
+            "text": text,
+            "parse_mode": "HTML",
+        }).encode()
+
+        for attempt in range(1, 4):
+            try:
+                with urllib.request.urlopen(url, data=data, timeout=10) as resp:
+                    result = json.loads(resp.read())
+                    if result.get("ok"):
+                        return True
+            except Exception as e:
+                log.warning("Telegram Fehler (Versuch %d/3): %s", attempt, e)
+                if attempt < 3:
+                    time.sleep(attempt * 2)
+
+        log.error("Telegram: Alert konnte nicht gesendet werden!")
+        return False