184 lines
7.5 KiB
Python
184 lines
7.5 KiB
Python
import paramiko
|
||
import time
|
||
import re
|
||
from collections import defaultdict
|
||
|
||
# Глобальный словарь для хранения предыдущих значений RX/TX (для расчета за час)
|
||
# Ключ: (host_ip, interface) -> {'rx': int, 'tx': int, 'time': timestamp}
|
||
prev_net_stats = {}
|
||
|
||
def get_ssh_client(host_config):
|
||
"""Создает SSH подключение"""
|
||
client = paramiko.SSHClient()
|
||
client.set_missing_host_key_policy(paramiko.AutoAddPolicy())
|
||
try:
|
||
client.connect(
|
||
hostname=host_config['ip'],
|
||
port=host_config['port'],
|
||
username=host_config['user'],
|
||
key_filename=host_config['key_path'],
|
||
timeout=10
|
||
)
|
||
return client
|
||
except Exception as e:
|
||
return f"SSH Error: {str(e)}"
|
||
|
||
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 collect_metrics(host_config):
|
||
"""Сбор всех метрик с одного хоста"""
|
||
client = get_ssh_client(host_config)
|
||
if isinstance(client, str):
|
||
return {"error": client}
|
||
|
||
result = {
|
||
"host": host_config['name'],
|
||
"ip": host_config['ip'],
|
||
"metrics": {}
|
||
}
|
||
|
||
# 1. 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
|
||
|
||
# 2. 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)
|
||
else:
|
||
result["metrics"]["mem_used_mb"] = 0
|
||
result["metrics"]["mem_percent"] = 0
|
||
|
||
# 3. Uptime и Load Average
|
||
uptime_output = exec_cmd(client, "uptime")
|
||
# Парсим load average (1,5,15 мин)
|
||
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)
|
||
# Парсим uptime (дни/часы)
|
||
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"
|
||
|
||
# 4. Сеть (трафик за последний час)
|
||
# Сначала собираем текущие счетчики по всем интерфейсам (кроме lo)
|
||
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}
|
||
|
||
# Расчет за час (используем глобальный кеш)
|
||
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:
|
||
# Считаем разницу. Если перезагрузили сервер - счетчики сбросятся, это ок.
|
||
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
|
||
|
||
# Переводим в Мегабайты (для удобства)
|
||
result["metrics"]["net_rx_mb"] = round(total_rx_hour / (1024 * 1024), 2)
|
||
result["metrics"]["net_tx_mb"] = round(total_tx_hour / (1024 * 1024), 2)
|
||
|
||
# Сохраняем текущие значения для следующего замера
|
||
prev_net_stats[host_key] = current_net
|
||
|
||
# 5. 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 = []
|
||
# Собираем все диски, показываем топ по нагрузке
|
||
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(',', '.')) # read/s
|
||
wps = float(parts[4].replace(',', '.')) # write/s
|
||
disk_stats.append({
|
||
"name": disk_name,
|
||
"util": util,
|
||
"rps": rps,
|
||
"wps": wps
|
||
})
|
||
except:
|
||
pass
|
||
|
||
# Сортируем по утилизации и берем топ-3 самых нагруженных
|
||
disk_stats.sort(key=lambda x: x['util'], reverse=True)
|
||
result["metrics"]["disks"] = disk_stats[:3]
|
||
|
||
# 6. Температуры (если есть sensors)
|
||
temp_output = exec_cmd(client, "sensors -u 2>/dev/null | grep -E 'temp[0-9]+_input|Core' | head -5")
|
||
temps = []
|
||
if temp_output and "Error" not in temp_output:
|
||
# Парсим цифры
|
||
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)))
|
||
result["metrics"]["temps"] = temps
|
||
|
||
# 7. 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
|
||
|
||
# 8. Логи (ошибки, fail2ban)
|
||
# Проверяем системный журнал на панику/ошибки за последнюю минуту (чтобы не спамить)
|
||
last_logs = exec_cmd(client, "journalctl -p 3 -xn 5 --no-pager 2>/dev/null || dmesg | tail -5")
|
||
# Fail2ban статус
|
||
fail2ban_status = exec_cmd(client, "fail2ban-client status 2>/dev/null | grep 'Number of failed'")
|
||
if not fail2ban_status:
|
||
fail2ban_status = "Fail2ban not installed or inactive"
|
||
|
||
# Кто пытался прорваться (последние записи из логов)
|
||
fail2ban_logs = exec_cmd(client, "tail -10 /var/log/fail2ban.log 2>/dev/null | grep -i banned || echo 'No recent bans'")
|
||
|
||
result["metrics"]["logs_errors"] = last_logs
|
||
result["metrics"]["fail2ban_status"] = fail2ban_status
|
||
result["metrics"]["fail2ban_logs"] = fail2ban_logs
|
||
|
||
client.close()
|
||
return result
|