network-probe

A probe for checking the network health and reporting the health to a remote InfluxDB via HTTP API.

Intended to be run on a Raspberry Pi connected to the target network. The software will ping predefined adresses inside and outside the network, with predefined freaquency, to catch connection drops. Also probes the default DNS to see if it has outages.

The target server is an InfluxDB. The data from multiple probes can be easily analyzed with Grafana, for example.

Compatibility:

Compatability Notice:

This version has legacy support for Python 3.5. This support, however will be discontinued in the future. In the future, you will need Python 3.6 or newer. Older versions of Raspbian, such as Raspbian 9 Stretch, don't have Python 3.6.

This version is designed to work with Raspbian Stretch and Python 3.5 and newer and Windows. Compatability with other Linux versions varies. It is dependent on the output of ping and traceroute commands.

Features

• pinging predefined addresses
• crude network detection
• logging to file
• Config with config.json
  • see: example_config.json

Install

Install Python 3 on Linux

$ sudo apt-get update
$ sudo apt-get install python3
$ sudo apt-get install python3-pip

Install dependencies

pip3 install -r requirements.txt

Dependencies

• Python 3.5* or newer
• InfluxDB Python HTTP API

Config

configure with config.json file

The config part is divided in to two parts.

Primary config

The first part has all the configuration you need to care about

First is the name of your probe and an id. These are used as tags in the data sent back to the InfluxDB.

"name": "aProbe",
"id": 0,

Targets is the list of targets you want to ping.

"targets": [
    "1.1.1.1",
    "1.0.0.1",
    "google.com"
],

Key	Default	Explanation
"time_interval"	2	Time between pings
"ping_count"	1	Times to ping per test
"detection_depth"	0	How meny hops from the probe are added to ping list
"event_queue"	15000	How meny pings are buffered if network is interrupted.
"db_name"	"db"	The InfluxDB name
"db_user"	"user"	Username to log in to the InfluxDB
"db_password"		the InfluxDB password
"db_host"	"localhost"	the address to the InfluxDB. !! omit 'https:\\' !!
"db_port"	8086	Port used to connect to the InfluxDB

detection_depth:

The probe can automatically ping devices on path to 1.1.1.1. These devices may include your router, your ISP and other network infrasturcture. Can be used for a quick and dirty shotgun approach.

Select the number of devices or "hops", counting from the probe, you want to ping.

Zero (0) disables this feature. Disabling is recommended for long deployments and multiple devices. The path may change between restarts and mess with your data. Different probes on the same network may get different paths which makes their data un-comparable. See also notes on pinging the internet.

Message template config

The second part contains a template for the message body.

YOU DO NOT NEED TO MODIFY IT

Infact, best you leave it alone, unless you know what you're doing.

  {
      "measurement": "ping",
      "tags": {
          "target": "",
          "name": "Test",
          "id": 0
      },
      "time": 0,
      "fields": {
          "up": 1
      }
  }

Run

python3 probe.py

stop with Ctrl + C

Run on boot (Raspbian Linux)

Disclaimer:

Included is a simple systemd service that automatically starts the probe on boot after network is up. Please note that both the installation script and the service itself have hardcoded paths for specific use case with Raspberry Pi.

Assuming

• the OS Raspbian Linux and
• user the pi (raspbian default user)

If you have not yet done so, clone network-probe to Pi home directory:

cd /home/pi
git clone https://github.com/JValtteri/network-probe.git

Run

$ install-service.sh

You can check the status, restart and stop the service with the following commands respectively.

$ sudo sytemctl status network-probe
$ sudo sytemctl restart network-probe
$ sudo sytemctl stop network-probe

Notes on Pinging the internet

Note that pinging the same device or IP on any network may be interpreted by some firewalls as a DOS attack. In our experience, even Cloudflares 1.1.1.1 will start to throttle our pings after a day or two of pinging at a total rate of ~3 pings/s, equivalent to six (6) devices on default settings, pinging from the same IP. Consider using internet root servers as targets for longer or larger deployments.