PART 03 — 運用・保守要件定義書
システム監視・バックアップ・デプロイ手順・インシデント対応・SLA管理

"""
複数エンドポイントの死活監視を行い、異常検知時にメール通知するスクリプト。
pip install requests

活用例: Cronで5分おきに実行し、2回連続失敗で通知する運用監視ツール
"""
import requests
import smtplib
import json
from pathlib import Path
from datetime import datetime
from dataclasses import dataclass, field
from email.mime.text import MIMEText
from email.mime.multipart import MIMEMultipart


# 監視設定
TARGETS = [
    {"name": "Webサーバー (本番)", "url": "https://example.com/health/", "timeout": 10},
    {"name": "API サーバー", "url": "https://api.example.com/api/v1/health/", "timeout": 10},
    {"name": "管理画面", "url": "https://admin.example.com/health/", "timeout": 10},
]

ALERT_EMAIL = "ops-team@example.com"
SMTP_HOST = "smtp.example.com"
SMTP_PORT = 587
SMTP_USER = "monitor@example.com"
SMTP_PASS = "your-smtp-password"

# 連続失敗カウント保存ファイル
STATE_FILE = Path("/var/monitor/health_state.json")


@dataclass
class CheckResult:
    name: str
    url: str
    success: bool
    status_code: int = 0
    response_time_ms: float = 0.0
    error: str = ""


def check_health(target: dict) -> CheckResult:
    """エンドポイントにGETリクエストを送りヘルスチェックを行う"""
    try:
        start = datetime.now()
        resp = requests.get(target["url"], timeout=target["timeout"], allow_redirects=False)
        elapsed_ms = (datetime.now() - start).total_seconds() * 1000
        success = (200 <= resp.status_code < 300)
        return CheckResult(
            name=target["name"], url=target["url"],
            success=success, status_code=resp.status_code,
            response_time_ms=elapsed_ms,
            error="" if success else f"HTTP {resp.status_code}"
        )
    except requests.Timeout:
        return CheckResult(name=target["name"], url=target["url"], success=False, error="タイムアウト")
    except Exception as e:
        return CheckResult(name=target["name"], url=target["url"], success=False, error=str(e))


def load_state() -> dict:
    if STATE_FILE.exists():
        return json.loads(STATE_FILE.read_text())
    return {}


def save_state(state: dict) -> None:
    STATE_FILE.parent.mkdir(parents=True, exist_ok=True)
    STATE_FILE.write_text(json.dumps(state, indent=2, ensure_ascii=False))


def send_alert(subject: str, body: str) -> None:
    """管理者にアラートメールを送信する"""
    msg = MIMEMultipart()
    msg["From"] = SMTP_USER
    msg["To"] = ALERT_EMAIL
    msg["Subject"] = subject
    msg.attach(MIMEText(body, "plain", "utf-8"))
    with smtplib.SMTP(SMTP_HOST, SMTP_PORT) as server:
        server.starttls()
        server.login(SMTP_USER, SMTP_PASS)
        server.send_message(msg)


def run_health_checks() -> None:
    state = load_state()
    now_str = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
    alerts = []

    for target in TARGETS:
        result = check_health(target)
        key = target["name"]
        prev = state.get(key, {"fail_count": 0, "alerted": False})

        if result.success:
            if prev["alerted"]:
                # 復旧通知
                send_alert(
                    f"[RESOLVED] {key} が復旧しました",
                    f"システム: {key}\nURL: {result.url}\n復旧時刻: {now_str}"
                )
            state[key] = {"fail_count": 0, "alerted": False}
            status = f"✅ {key}: OK ({result.response_time_ms:.0f}ms)"
        else:
            fail_count = prev["fail_count"] + 1
            alerted = prev["alerted"]
            state[key] = {"fail_count": fail_count, "alerted": alerted}

            if fail_count >= 2 and not alerted:
                # 2回連続失敗で初回アラート
                alerts.append(f"❌ {key}: {result.error} ({fail_count}回連続失敗)")
                state[key]["alerted"] = True

            status = f"❌ {key}: {result.error} (失敗{fail_count}回)"

        print(f"[{now_str}] {status}")

    if alerts:
        body = "\n".join(alerts)
        body += f"\n\n確認時刻: {now_str}"
        send_alert(f"[ALERT] {len(alerts)}件のエンドポイントが応答しません", body)
        print(f"⚠️ アラート送信: {len(alerts)}件")

    save_state(state)


if __name__ == "__main__":
    run_health_checks()