Forward Zabbix Alerts to WeChat via Kafka – Complete Step‑by‑Step Guide

Application scenario Because the target company has strict security requirements, the Zabbix server cannot access the Internet, so alerts cannot be sent directly to instant‑messaging tools. The solution is to forward alerts to a middleware—Kafka—and let another component push them to the desired channel.

Basic environment A Kafka cluster is already deployed (installation details omitted).

<code>vim /etc/hosts
192.168.179.133 kafka3
192.168.179.132 kafka2
192.168.179.131 kafka1</code>

Start Zookeeper and Kafka To allow consumers to evenly consume data, a topic with 6 partitions and 3 replicas is created.

<code># Start Zookeeper
cd /root/kafka_2.12-2.4.1
nohup bin/zookeeper-server-start.sh config/zookeeper.properties &>> zookeeper.log &
# Start Kafka
nohup bin/kafka-server-start.sh config/server.properties &>> kafka.log &
# Create topic
bin/kafka-topics.sh --create --zookeeper kafka1:2181,kafka2:2181,kafka3:2181 \
    --replication-factor 3 --partitions 6 --topic zabbix-alert
# List topics
bin/kafka-topics.sh --list --bootstrap-server 192.168.179.132:9092</code>

Script to send Zabbix alerts to Kafka

<code>vim /usr/lib/zabbix/alertscripts/alert_kafka.py
chmod +x /usr/lib/zabbix/alertscripts/alert_kafka.py
#!/usr/bin/python
#coding=utf-8
from kafka import KafkaProducer
import json, sys
receive = sys.argv[1]
message = sys.argv[2]
producer = KafkaProducer(
    value_serializer=lambda v: json.dumps(v).encode('utf-8'),
    bootstrap_servers=['192.168.179.132:9092','192.168.179.133:9092','192.168.179.134:9092']
)
data = {"receive": receive, "alert": message}
producer.send('zabbix-alert', data)
producer.close()</code>

Script to consume alerts from Kafka and push to Enterprise WeChat

<code>vim /usr/lib/zabbix/alertscripts/receive.py
#!/usr/bin/python
#coding=utf-8
from kafka import KafkaConsumer
import json, requests, sys
from requests.packages.urllib3.exceptions import InsecureRequestWarning
requests.packages.urllib3.disable_warnings(InsecureRequestWarning)
reload(sys)
sys.setdefaultencoding('utf-8')
corpid = 'ww95f3a6ffd22fiyhe8b8'
secret = 'iJjHQxTCjIMyW7ZjzityucxqA7Hg2fjcLM2ssdvwY1Zc'
agentid = '1000002'

# Consume messages from the zabbix-alert topic
consumer = KafkaConsumer('zabbix-alert', group_id='zabbix-alert',
    bootstrap_servers=['192.168.179.132:9092','192.168.179.133:9092','192.168.179.134:9092'],
    auto_offset_reset='earliest', value_deserializer=json.loads)

def gettoken():
    tokenurl = "https://qyapi.weixin.qq.com/cgi-bin/gettoken"
    data = {"corpid": corpid, "corpsecret": secret}
    r = requests.get(url=tokenurl, params=data, verify=False)
    return r.json()['access_token']

def sendweixin(token, messages):
    wechaturl = "https://qyapi.weixin.qq.com/cgi-bin/message/send?access_token=%s" % token
    data = {
        "touser": messages['receive'],
        "msgtype": "text",
        "agentid": agentid,
        "text": {"content": messages['alert']},
        "safe": "0"
    }
    headers = {'content-type': 'application/json'}
    req = requests.post(url=wechaturl, headers=headers, json=data, verify=False)
    print(req.text)
    return req

for message in consumer:
    token = gettoken()
    messages = message.value
    sendweixin(token, messages)</code>

Run the consumer script in the background:

<code>nohup python alert_receive.py &>> alert.log &</code>

Zabbix action configuration

Create alert media

Configure user media

Set the recipient to the Enterprise WeChat user ID.

Configure action

Test the alert after configuration

The resulting notification appears in Enterprise WeChat as shown below.