网络设备异常还在人工排查?用 Python 实现批量自动诊断 + 一键生成诊断报告(一)(附源码)

[

  {

    "name": "shwh_hexin",

    "device_type": "hp_comware",

    "host": "192.168.31.253",

    "username": "admin",

    "password": "你的密码"

  },

  {

    "name": "shwh02",

    "device_type": "hp_comware",

    "host": "192.168.31.252",

    "username": "admin",

    "password": "你的密码"

  }

]

# -*- coding: utf-8 -*-

import json

from netmiko import ConnectHandler

import pandas as pd

import re

import numpy as np

from openpyxl.styles import PatternFill

from openpyxl import load_workbook

# ======================

# 读取设备

# ======================

with open("devices.json", "r", encoding="utf-8") as f:

    devices = json.load(f)

# ======================

# 连接设备

# ======================

def connect_device(device):

    try:

        print(f"\n🔍 连接 {device['name']}...")

        conn = ConnectHandler(

            device_type="hp_comware",

            host=device['host'],

            username=device['username'],

            password=device['password'],

            timeout=10,

            global_delay_factor=2

        )

        print("✅ 连接成功")

        return conn

    except Exception as e:

        print(f"❌ 连接失败: {e}")

        return None

# ======================

# 数据采集

# ======================

def collect_data(conn):

    return {

        "intf": conn.send_command("display interface", read_timeout=20),

        "cpu": conn.send_command("display cpu-usage"),

        "mem": conn.send_command("display memory summary"),

        "lag": conn.send_command("display link-aggregation verbose"),

        "log": conn.send_command("display logbuffer")

    }

# ======================

# 解析函数

# ======================

def parse_cpu(cpu):

    values = [int(x) for x in re.findall(r"(\d+)%", cpu)]

    return max(values) if values else None

def parse_mem(mem):

    match = re.findall(r"(\d+\.\d+)%", mem)

    if match:

        free_ratio = float(match[-1])

        return round(100 - free_ratio, 1)

    return None

def parse_flap(log):

    return re.findall(r"GigabitEthernet\d+/\d+/\d+", log)

def parse_lag(lag):

    lags = []

    blocks = lag.split("Bridge-Aggregation")

    for block in blocks[1:]:

        name = "Bridge-Aggregation" + block.split()[0]

        state = "DOWN" if "Current state: DOWN" in block else "UP"

        lags.append(f"{name}({state})")

    return lags

def parse_interface_traffic(intf):

    result = {}

    current_intf = None

    for line in intf.splitlines():

        m = re.match(r"(GigabitEthernet\d+/\d+/\d+)", line)

        if m:

            current_intf = m.group(1)

        t = re.search(r"(\d+)\s+bytes/sec", line)

        if t and current_intf:

            result[current_intf] = int(t.group(1))

    return result

def format_traffic(val):

    if val >= 1_000_000:

        return f"{round(val/1_000_000,2)}MB/s"

    elif val >= 1_000:

        return f"{round(val/1_000,2)}KB/s"

    else:

        return f"{val}B/s"

def detect_traffic_abnormal(intf_traffic):

    abnormal = []

    if not intf_traffic:

        return abnormal

    values = list(intf_traffic.values())

    mean = np.mean(values)

    std = np.std(values)

    for intf, val in intf_traffic.items():

        if val > mean + 2 * std and val > 100000:

            abnormal.append(f"{intf}({format_traffic(val)})")

    return abnormal

# ======================

# 诊断

# ======================

def diagnose(data):

    cpu = parse_cpu(data['cpu'])

    mem = parse_mem(data['mem'])

    flap = parse_flap(data['log'])

    lag = parse_lag(data['lag'])

    intf_traffic = parse_interface_traffic(data['intf'])

    traffic_abnormal = detect_traffic_abnormal(intf_traffic)

    lag_problem = [x for x in lag if "DOWN" in x]

    return {

        "接口flap": ",".join(set(flap)) if flap else "正常",

        "CPU": f"{cpu}%" if cpu else "未获取",

        "内存": f"{mem}%" if mem else "未获取",

        "聚合口": ",".join(lag) if lag else "未配置",

        "流量异常": ",".join(traffic_abnormal) if traffic_abnormal else "正常",

        "总问题": "异常" if flap or lag_problem or traffic_abnormal else "正常"

    }

# ======================

# 主程序

# ======================

results = []

for dev in devices:

    conn = connect_device(dev)

    if conn:

        data = collect_data(conn)

        conn.disconnect()

        res = diagnose(data)

    else:

        res = {k: "连接失败" for k in ["接口flap","CPU","内存","聚合口","流量异常","总问题"]}

    res["设备"] = dev['name']

    results.append(res)

# ======================

# 写入Excel

# ======================

df = pd.DataFrame(results)

file = "network_report.xlsx"

df.to_excel(file, index=False)

# ======================

# 精准高亮（只标异常字段）

# ======================

wb = load_workbook(file)

ws = wb.active

red = PatternFill("solid", fgColor="FF0000")

for r in range(2, ws.max_row + 1):

    # 接口flap

    if ws.cell(r, 1).value != "正常":

        ws.cell(r, 1).fill = red

    # CPU > 80%

    cpu = ws.cell(r, 2).value

    if cpu and cpu != "未获取":

        if int(cpu.replace("%", "")) > 80:

            ws.cell(r, 2).fill = red

    # 内存 > 80%

    mem = ws.cell(r, 3).value

    if mem and mem != "未获取":

        if float(mem.replace("%", "")) > 80:

            ws.cell(r, 3).fill = red

    # 聚合口 DOWN

    lag = ws.cell(r, 4).value

    if lag and "DOWN" in lag:

        ws.cell(r, 4).fill = red

    # 流量异常

    if ws.cell(r, 5).value != "正常":

        ws.cell(r, 5).fill = red

    # 总问题

    if ws.cell(r, 6).value != "正常":

        ws.cell(r, 6).fill = red

wb.save(file)

print("\n🎉 完成！生成 network_report.xlsx（已优化高亮）")