Python 3- Send Data to Zabbix Server

Following is method to send data to zabbix (2.4.6)  server written in python 3.

Corresponding Item key should be configured as zabbix trapper on zabbix server to successfully process this request.

	def send_zabbix_request(self, data_json):
	HEADER = "ZBXD\1"
	data_json_str = json.dumps(data_json)
	logger.debug("Zabbix Request ----- " + data_json_str)
	data_len= len(data_json_str)
	data_header = struct.pack('<Q', data_len)
	zabbix_req = HEADER.encode() + data_header + data_json_str.encode()
	try:
	sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
	sock.connect ((self.zabbix_host, self.zabbix_port))
	sock.send(zabbix_req)
	except Exception as err:
	logger.critical("Error connecting to Zabbix Server -- {0}" .format( err))
	return json_err("Error connecting to Zabbix Server {0}" .format( err))
	else:
	response_header = sock.recv(5).decode()
	if not response_header == 'ZBXD\1':
	logger.critical('Got invalid responsei from Zabbix')
	return json_err('Got invalid responsei from Zabbix')
	response_data_header = sock.recv(8)
	response_len = struct.unpack('<Q', response_data_header)[0]
	response_raw = sock.recv(response_len)
	sock.close()
	return json.loads(response_raw.decode())

view raw send_zabbix_request.py hosted with ❤ by GitHub

Zabbix data is a json in following format.

{
  "request":"sender data",
  "data":[
     {
       "host":"hostName",
       "key":"key-1",
       "value":"200",
       "clock":1458748060
     },
     {
       "host":"hostName",
       "key":"key-4",
       "value":"test",
       "clock":1458748060}
     ]}
}

SQL Query- Where Value Between Two Columns

Imagine a simplified sample table that contains id, low_range, high_range. Low_range and High_range can be decimal ip address, zip codes or even timestamps.

id	low_range	high_range
1	400002	400120
5	407049	407340
5003	637190	637805
702984	849380	849875

We need to find row where a number is greater than low_range and lesser than high_range. Assumption is that maximum of one record and minimum of zero will be returned. 

SELECT id FROM table WHERE num > low_range AND num < high_range. 

The sql above might work well when we have probably few hundreds rows. But it is extremely slow with just few thousands records. Adding index on min_range (or / and max_range) will be helpful only for last few (or first few) records in index. For rest of rows, engine will consider a full table scan to be faster compared to using index. 

Solution 1:

SELECT id
FROM tab tab,
( SELECT max(low_range) as max_low_range
           FROM tab WHERE low_range <= num ) low_range_table
WHERE table.low_range = low_range_table.max_low_range
AND full_table.high_range >= num;

While above query might look complex to human, it makes life easier for database engine.   It first finds max low range available lesser than number. We have index on low_range; so this sub query will be fast. Next it joins with full table using this low_range. (Ideally, we should join with primary key). This will give exactly one row. Finally it validates if high range is still higher than number / string.

Explain plan:

Solution 2: 

          SELECT id
          FROM (
                    SELECT * FROM tab tab
                    WHERE low_range <= num
                     ORDER BY low_range DESC LIMIT 1) as max_low_range
          WHERE max_low_range.high_range >= num

Explain Plan