Automate Zabbix Monitoring: Fetch Host Metrics and Export to CSV with Python

Implementation Approach

Mainly using Zabbix's API to fetch and process data.

Zabbix provides a rich API to obtain host information, item IDs, and trend and history data.

First retrieve all hosts in a host group by group ID, including host name and IP address.

Iterate over each host ID and, within the loop, request the item ID based on the monitoring item key.

Use the obtained item IDs to fetch historical data and trend data respectively.

Store the history and trend values in a dictionary, then add each dictionary to a list.

Write the list information to an Excel/CSV file and schedule the script with a timed task.

Define Time Interval for Data Retrieval

<code>x=(datetime.datetime.now()-datetime.timedelta(minutes=120)).strftime("%Y-%m-%d %H:%M:%S")
y=(datetime.datetime.now()).strftime("%Y-%m-%d %H:%M:%S")
def timestamp(x,y):
    p=time.strptime(x,"%Y-%m-%d %H:%M:%S")
    starttime = str(int(time.mktime(p)))
    q=time.strptime(y,"%Y-%m-%d %H:%M:%S")
    endtime= str(int(time.mktime(q)))
    return starttime,endtime</code>

Get Host Information by Host Group ID

<code>def get_hosts(groupids,auth):
    data ={
            "jsonrpc": "2.0",
             "method": "host.get",
             "params": {
             "output": [ "name"],
             "groupids": groupids,
             "filter":{
                 "status": "0"
             },
             "selectInterfaces": [   
                        "ip"
                    ],
            },
            "auth": auth,  # the auth id is what auth script returns, remember it is string
            "id": 1
        }
    gethost=requests.post(url=ApiUrl,headers=header,json=data)
    return json.loads(gethost.content)["result"]</code>

Build Loop Using Host Info to Retrieve Monitoring Item Data

Fetch Historical Data

<code>host=[]
    print(hosts)
    for i in hosts:
        item1=[]
        item2=[]
        #print(i)
        dic1={}
        for j in ['vfs.fs.size[C:,total]','vm.memory.size[total]','system.cpu.num']:
            data={
                "jsonrpc": "2.0",
                "method": "item.get",
                "params": {
                    "output": [
                        "itemid"

                    ],
                    "search": {
                        "key_": j  
                    },
                    "hostids": i['hostid']
                },
                "auth":auth,
                "id": 1
            }
            getitem=requests.post(url=ApiUrl,headers=header,json=data)
            item=json.loads(getitem.content)['result']

            hisdata={
                "jsonrpc":"2.0",
                "method":"history.get",
                "params":{
                    "output":"extend",                    
                    "time_from":timestamp[0],
                    #"time_till":timestamp[1],
                    "history":0,
                    "sortfield": "clock",
                    "sortorder": "DESC",
                    "itemids": '%s' %(item[0]['itemid']),
                    "limit":1
                },
                "auth": auth,
                "id":1
                }
            gethist=requests.post(url=ApiUrl,headers=header,json=hisdata)
            hist=json.loads(gethist.content)['result']
            item1.append(hist)</code>

Fetch Trend Data

<code>for j in ['vfs.fs.size[C:,used]','vm.memory.size[used]','system.cpu.load']:
            data={
                "jsonrpc": "2.0",
                "method": "item.get",
                "params": {
                    "output": [
                        "itemid"

                    ],
                    "search": {
                        "key_": j  
                    },
                    "hostids": i['hostid']
                },
                "auth":auth,
                "id": 1
            }
            getitem=requests.post(url=ApiUrl,headers=header,json=data)
            item=json.loads(getitem.content)['result']

            trendata={
                "jsonrpc":"2.0",
                "method":"trend.get",
                "params":{
                    "output": [
                        "itemid",
                        "value_max",
                        "value_avg"
                    ],                    
                    "time_from":timestamp[0],
                    "time_till":timestamp[1],
                    "itemids": '%s' %(item[0]['itemid']),
                    "limit":1
                },
                "auth": auth,
                "id":1
                }
            gettrend=requests.post(url=ApiUrl,headers=header,json=trendata)
            trend=json.loads(gettrend.content)['result']
            item2.append(trend)</code>

Process Retrieved Data and Export to CSV

<code>dic1['Hostname']=i['name']
        dic1['IP']=i['interfaces'][0]['ip']
        dic1['磁盘C:Total(B)']=round(float(item1[0][0]['value'])/1024**3,2)
        dic1['磁盘最大C:Used(B)']=round(float(item2[0][0]['value_max'])/1024**3,2)
        dic1['内存Total(B)']=round(float(item1[1][0]['value'])/1024**3,2)
        dic1['内存最大Used(B)']=round(float(item2[1][0]['value_max'])/1024**3,2)
        dic1['内存平均used(B)']=round(float(item2[1][0]['value_avg'])/1024**3,2)
        dic1['CPU负载最大值']=item2[2][0]['value_max']
        dic1['CPU负载平均值']=item2[2][0]['value_avg']
        dic1['CPU 核数']=item1[2][0]['value']
        x = time.localtime(int(item1[2][0]['clock']))
        item1[2][0]['clock'] = time.strftime("%Y-%m-%d %H:%M:%S", x)
        dic1['clock']=item1[2][0]['clock']
        host.append(dic1)  
        print(item)
    print(host)
    return host       
def writecsv(getitem1):
    with open('data.csv','w',encoding='utf-8-sig') as f:
        #f.write(codecs.BOM_UTF8)
        writer = csv.DictWriter(f,csvheader)
        writer.writeheader()
        for row in getitem1:
            writer.writerow(row)</code>

Resulting Output

Full code is available on GitHub: https://github.com/sunsharing-note/zabbix/blob/master/xunjian_auto.py

Zabbix API documentation: https://www.zabbix.com/documentation/4.0/zh/manual/api/reference/history/get

Feel free to discuss.