Run a custom command using consul hashicorp json file

Question

i set up a 3 consul servers and 1 nginx and linked them in one cluster and everything is ok on the web ui .

i would like to run a custom .JSON config file to run this command "ss -nlt" in order to see the number of connections to that server on the web .

i already done similar thing for nginx as the below JSON file :

  {
        "service": {
                "name": "Nginx1-server",
                "tags": [ "colourserver" ],
                "port": 80,
                "check": {
                        "id": "webserver_up_test",
                        "name": "Get Nginx Main Page",
                        "http": "http://localhost/index.html",
                        "interval": "3s",
                        "timeout": "1s"
                }
        }
}

any ideas ? or is that even possible ?

Answer 1

I figured it out , i used a python script and called it from a json file and it worked !

this is my python file "test.py" :

#! /usr/bin/env python3
import os, sys
cmd = "ss -nlt | grep 'LISTEN' -c"
print ("Hello Mahdi , This is the number of connections right now" )
os.system(cmd)

and this is my json file :

{
        "service": {
                "name": "Nginx1-server",
                "tags": [ "colourserver" ],
                "port": 80,
                "check": {
                        "id": "Connections",
                        "name": "Number of Connections",
                        "args": ["usr/local/bin/test.py"],
                        "interval": "10s",
                        "timeout": "1s"
                }
}
}

and the results are :

Answer 2

A more reliable way to obtain the number of active connections to the web server would be to use nginx's stub_status module. This module outputs the number of active connections being handled by the web server. Below is an example config:

consul-config.json

{
  "enable_script_checks": false,
  "enable_local_script_checks": true,
  "service": {
    "name": "Nginx1-server",
    "tags": [
      "colourserver"
    ],
    "port": 80,
    "check": {
      "id": "Connections",
      "name": "Number of Connections",
      "args": [
        "/usr/local/bin/nginx-connections.sh"
      ],
      "interval": "10s",
      "timeout": "1s"
    }
  }
}

nginx-connections.sh

#!/usr/bin/env bash

curl --no-progress-meter http://127.0.0.1:8080/status | awk '/^Active/ { print $0 }'

nginx.conf

worker_rlimit_nofile 8192;

events {
}

http {
  server {
    listen       8080;
    server_name  _;
    root         /var/www/html;

    location /status {
        stub_status;
    }
  }
}

The number of active connections can then be seen in the output from the script check.

$ curl --get 127.0.0.1:8500/v1/agent/checks --data 'filter=CheckID=="Connections"'
{
  "Connections": {
    "Node": "blake-C02YX6QSLVCG",
    "CheckID": "Connections",
    "Name": "Number of Connections",
    "Status": "passing",
    "Notes": "",
    "Output": "Active connections: 1 \n",
    "ServiceID": "Nginx1-server",
    "ServiceName": "Nginx1-server",
    "ServiceTags": [
      "colourserver"
    ],
    "Type": "script",
    "Interval": "10s",
    "Timeout": "10s",
    "ExposedPort": 0,
    "Definition": {},
    "CreateIndex": 0,
    "ModifyIndex": 0
  }
}