tg_monitor_bot/monitor.py

334 lines
14 KiB
Python
Raw Normal View History

2026-08-02 21:46:32 +00:00
import paramiko
import time
import re
import os
from collections import defaultdict
import logging
logger = logging.getLogger(__name__)
# Глобальный словарь для хранения предыдущих значений RX/TX
prev_net_stats = {}
def get_ssh_client(host_config):
"""Создает SSH подключение через ssh-agent или ключ"""
client = paramiko.SSHClient()
client.set_missing_host_key_policy(paramiko.AutoAddPolicy())
try:
logger.info(f"🔌 Подключаюсь к {host_config['ip']}:{host_config['port']} как {host_config['user']}")
# Сначала пробуем ssh-agent
try:
from paramiko.agent import Agent
agent = Agent()
if agent.get_keys():
logger.info(f"🔑 Использую ssh-agent для {host_config['ip']}")
client.connect(
hostname=host_config['ip'],
port=host_config['port'],
username=host_config['user'],
timeout=10,
allow_agent=True,
look_for_keys=True
)
logger.info(f"✅ Подключено к {host_config['ip']} через ssh-agent")
return client
except Exception as e:
logger.debug(f"ssh-agent не доступен: {e}")
# Если ssh-agent не работает, пробуем через ключ
if os.path.exists(host_config['key_path']):
logger.info(f"🔑 Использую ключ: {host_config['key_path']}")
client.connect(
hostname=host_config['ip'],
port=host_config['port'],
username=host_config['user'],
key_filename=host_config['key_path'],
timeout=10
)
logger.info(f"✅ Подключено к {host_config['ip']} по ключу")
return client
else:
return f"❌ Файл ключа не найден: {host_config['key_path']}"
except paramiko.AuthenticationException:
error_msg = f"❌ Ошибка аутентификации для {host_config['ip']}! Проверь логин, ключ или ssh-agent."
logger.error(error_msg)
return error_msg
except paramiko.SSHException as e:
error_msg = f"❌ SSH ошибка для {host_config['ip']}: {str(e)}"
logger.error(error_msg)
return error_msg
except Exception as e:
error_msg = f"❌ Неизвестная ошибка для {host_config['ip']}: {str(e)}"
logger.error(error_msg)
return error_msg
def exec_cmd(client, command):
"""Выполняет команду и возвращает вывод"""
try:
stdin, stdout, stderr = client.exec_command(command, timeout=10)
return stdout.read().decode('utf-8').strip()
except Exception as e:
return f"CMD Error: {str(e)}"
def detect_host_type(client):
"""Определяет тип хоста: physical, vm, lxc, container"""
# Проверка на LXC
lxc_check = exec_cmd(client, "cat /proc/1/environ 2>/dev/null | tr '\\0' '\\n' | grep -i lxc || echo ''")
if lxc_check:
return "lxc"
# Проверка на контейнер (Docker и др)
container_check = exec_cmd(client, "cat /proc/1/environ 2>/dev/null | tr '\\0' '\\n' | grep -i container || echo ''")
if container_check:
return "container"
# Проверка на виртуальную машину
virt_check = exec_cmd(client, "systemd-detect-virt 2>/dev/null || echo 'none'")
if virt_check and virt_check not in ['none', '']:
return "vm"
# Проверка через dmidecode (если есть)
virt_check2 = exec_cmd(client, "dmidecode -s system-manufacturer 2>/dev/null || echo 'Unknown'")
if virt_check2:
virt_vendors = ['vmware', 'virtualbox', 'kvm', 'qemu', 'xen', 'microsoft', 'parallels']
if any(vendor in virt_check2.lower() for vendor in virt_vendors):
return "vm"
# Если ничего не найдено - считаем физическим железом
return "physical"
def check_alerts(metrics, host_type):
"""Проверяет критические метрики и возвращает список алертов"""
alerts = []
# CPU алерт (>80%)
cpu = metrics.get('cpu', 0)
if cpu > 80:
alerts.append(f"🔴 ВЫСОКАЯ ЗАГРУЗКА CPU: {cpu:.1f}%")
elif cpu > 60:
alerts.append(f"🟡 ПОВЫШЕННАЯ ЗАГРУЗКА CPU: {cpu:.1f}%")
# RAM алерт (>85%)
mem_percent = metrics.get('mem_percent', 0)
if mem_percent > 85:
alerts.append(f"🔴 НЕХВАТКА ОЗУ: {mem_percent:.1f}% (использовано {metrics.get('mem_used_mb', 0)} MB)")
elif mem_percent > 70:
alerts.append(f"🟡 ПОВЫШЕННОЕ ПОТРЕБЛЕНИЕ ОЗУ: {mem_percent:.1f}%")
# Диск алерт (>80%)
disk_usage = metrics.get('disk_usage', 0)
if disk_usage > 80:
alerts.append(f"🔴 МАЛО МЕСТА НА ДИСКЕ: {disk_usage:.1f}%")
elif disk_usage > 70:
alerts.append(f"🟡 ЗАКАНЧИВАЕТСЯ МЕСТО НА ДИСКЕ: {disk_usage:.1f}%")
# OOM алерт
oom = metrics.get('oom', '')
if oom and "Нет OOM" not in oom:
alerts.append(f"💀 OOM KILLER СРАБОТАЛ! Проверь логи!")
# Load Average алерт (если > количество ядер)
load_1 = float(metrics.get('load_1', 0))
if load_1 > 4:
alerts.append(f"🔴 ВЫСОКИЙ LOAD AVERAGE: {load_1}")
elif load_1 > 2:
alerts.append(f"🟡 ПОВЫШЕННЫЙ LOAD AVERAGE: {load_1}")
return alerts
def collect_metrics(host_config):
"""Сбор всех метрик с одного хоста"""
client = get_ssh_client(host_config)
if isinstance(client, str):
return {"error": client, "host": host_config['name'], "ip": host_config['ip']}
result = {
"host": host_config['name'],
"ip": host_config['ip'],
"metrics": {}
}
# 1. Определяем тип хоста
host_type = detect_host_type(client)
result["host_type"] = host_type
result["metrics"]["host_type"] = host_type
# 2. CPU загрузка (%)
cpu_output = exec_cmd(client, "top -bn1 | grep 'Cpu(s)' | awk '{print $2}' | cut -d'%' -f1")
try:
result["metrics"]["cpu"] = float(cpu_output.replace(',', '.'))
except:
result["metrics"]["cpu"] = 0.0
# 3. ОЗУ (используется в MB и %)
mem_output = exec_cmd(client, "free -m | grep Mem")
parts = re.split(r'\s+', mem_output)
if len(parts) >= 3:
total_mem = int(parts[1])
used_mem = int(parts[2])
result["metrics"]["mem_used_mb"] = used_mem
result["metrics"]["mem_percent"] = round((used_mem / total_mem) * 100, 1)
result["metrics"]["mem_total_mb"] = total_mem
else:
result["metrics"]["mem_used_mb"] = 0
result["metrics"]["mem_percent"] = 0
result["metrics"]["mem_total_mb"] = 0
# 4. Uptime и Load Average
uptime_output = exec_cmd(client, "uptime")
load_avg = re.search(r'load average:\s+([\d.]+),\s+([\d.]+),\s+([\d.]+)', uptime_output)
if load_avg:
result["metrics"]["load_1"] = load_avg.group(1)
result["metrics"]["load_5"] = load_avg.group(2)
result["metrics"]["load_15"] = load_avg.group(3)
uptime_seconds = exec_cmd(client, "cat /proc/uptime | awk '{print $1}'")
try:
uptime_seconds = float(uptime_seconds.split('.')[0])
days = uptime_seconds // 86400
hours = (uptime_seconds % 86400) // 3600
minutes = (uptime_seconds % 3600) // 60
result["metrics"]["uptime"] = f"{int(days)}d {int(hours)}h {int(minutes)}m"
except:
result["metrics"]["uptime"] = "N/A"
# 5. Сеть (трафик за последний час)
net_output = exec_cmd(client, "cat /proc/net/dev | tail -n +3 | grep -v lo")
current_net = {}
for line in net_output.split('\n'):
if not line.strip():
continue
parts = re.split(r'\s+', line.strip())
iface = parts[0].replace(':', '')
rx_bytes = int(parts[1])
tx_bytes = int(parts[9])
current_net[iface] = {'rx': rx_bytes, 'tx': tx_bytes}
global prev_net_stats
host_key = host_config['ip']
total_rx_hour = 0
total_tx_hour = 0
if host_key in prev_net_stats:
old_data = prev_net_stats[host_key]
for iface, vals in current_net.items():
if iface in old_data:
delta_rx = vals['rx'] - old_data[iface]['rx']
delta_tx = vals['tx'] - old_data[iface]['tx']
if delta_rx > 0:
total_rx_hour += delta_rx
if delta_tx > 0:
total_tx_hour += delta_tx
result["metrics"]["net_rx_mb"] = round(total_rx_hour / (1024 * 1024), 2)
result["metrics"]["net_tx_mb"] = round(total_tx_hour / (1024 * 1024), 2)
prev_net_stats[host_key] = current_net
# 6. Диски (iostat %util)
iostat_output = exec_cmd(client, "iostat -x -k 1 2 | grep -E 'sd[a-z]|vd[a-z]|nvme' | tail -n +3")
disk_stats = []
for line in iostat_output.split('\n'):
if not line.strip():
continue
parts = re.split(r'\s+', line.strip())
if len(parts) >= 14:
disk_name = parts[0]
try:
util = float(parts[-1].replace(',', '.'))
rps = float(parts[3].replace(',', '.'))
wps = float(parts[4].replace(',', '.'))
disk_stats.append({
"name": disk_name,
"util": util,
"rps": rps,
"wps": wps
})
except:
pass
disk_stats.sort(key=lambda x: x['util'], reverse=True)
result["metrics"]["disks"] = disk_stats[:3]
# 7. Свободное место на диске (%)
disk_usage_output = exec_cmd(client, "df -h / | tail -1 | awk '{print $5}' | sed 's/%//'")
try:
result["metrics"]["disk_usage"] = float(disk_usage_output.replace(',', '.'))
except:
result["metrics"]["disk_usage"] = 0
# 8. Температуры (ТОЛЬКО ДЛЯ ФИЗИЧЕСКИХ ХОСТОВ!)
temps = []
if host_type == "physical":
temp_output = exec_cmd(client, "sensors -u 2>/dev/null | grep -E 'temp[0-9]+_input|Core' | head -5")
if temp_output and "Error" not in temp_output and "command not found" not in temp_output:
for line in temp_output.split('\n'):
if '_input:' in line:
val = re.search(r'([\d.]+)', line)
if val:
temps.append(float(val.group(1)))
else:
# Если sensors не установлен - пробуем через /sys/class/thermal
temp_sys = exec_cmd(client, "cat /sys/class/thermal/thermal_zone*/temp 2>/dev/null | head -3")
if temp_sys and "No such file" not in temp_sys:
for line in temp_sys.split('\n'):
try:
temps.append(float(line) / 1000)
except:
pass
result["metrics"]["temps"] = temps
# 9. Топ-5 процессов по CPU и RAM
top_cpu = exec_cmd(client, "ps aux --sort=-%cpu | head -6 | tail -5")
top_mem = exec_cmd(client, "ps aux --sort=-%mem | head -6 | tail -5")
result["metrics"]["top_cpu"] = top_cpu
result["metrics"]["top_mem"] = top_mem
# 10. Логи (ошибки)
last_logs = exec_cmd(client, "journalctl -p 3 -xn 5 --no-pager 2>/dev/null || dmesg | tail -5")
result["metrics"]["logs_errors"] = last_logs
# 11. Fail2ban
fail2ban_status = exec_cmd(client, "fail2ban-client status 2>/dev/null")
if not fail2ban_status or "not found" in fail2ban_status:
ps_fail2ban = exec_cmd(client, "ps aux | grep -v grep | grep fail2ban-server")
if ps_fail2ban:
fail2ban_status = "✅ Fail2ban запущен"
else:
fail2ban_status = "❌ Fail2ban не установлен или не запущен"
else:
lines = fail2ban_status.split('\n')
jails_info = []
for line in lines:
if '|- Number of jail:' in line or '`- Number of jail:' in line:
jails_info.append(line.strip())
if 'Currently banned:' in line:
jails_info.append(line.strip())
if jails_info:
fail2ban_status = "✅ Fail2ban активен\n" + "\n".join(jails_info)
fail2ban_logs = exec_cmd(client, "tail -20 /var/log/fail2ban.log 2>/dev/null | grep -E 'BAN|UNBAN' | tail -5 || echo 'Нет записей о банах'")
active_bans = exec_cmd(client, "fail2ban-client status 2>/dev/null | grep -A 5 'Currently banned:'")
if active_bans and "Currently banned:" in active_bans:
fail2ban_logs += "\n\n🔒 Активные баны:\n" + active_bans
result["metrics"]["fail2ban_status"] = fail2ban_status
result["metrics"]["fail2ban_logs"] = fail2ban_logs
# 12. OOM
oom_check = exec_cmd(client, "dmesg | grep -i 'out of memory' | tail -3 || echo 'Нет OOM-событий'")
if oom_check and "Нет OOM-событий" not in oom_check:
result["metrics"]["oom"] = oom_check
else:
result["metrics"]["oom"] = "✅ Нет OOM-событий"
# 13. Проверка алертов
result["metrics"]["alerts"] = check_alerts(result["metrics"], host_type)
client.close()
return result