melchor629 / uptime_monitor

An uptime monitor for your net devices in home, in a IOT device

Uptime monitor

Known when your network devices are connected to your home's network with a good precission in time. Everything in Docker containers, and ARM friendly. Uses fing to observe the changes and a little parser that listens for changes in a file fing edits and stores the data in an influxdb measure and some extra data into a redis database. It is available a simple page that allows you to monitor quickly the uptime.

How it works?

Uses fing app to detect the devices available in the network and if they are connected or not. Using the right arguments, fing can write the scanning into a .csv file. And now the little python script (called fing-parser) comes into the game. This script listen for changes in a folder, and when a file is changed (and is a csv) stores all the information in the influxdb and redis databases.

To make it work, fing should start with this arguments fing -d true -n NETWORK_IP/NETWORK_MASK -o table,csv,/fing/SOME_NAME.csv. The -d argument is to force obtain the domain names of every device. If you don't want that, you can disable with a false. The NETWORK_IP/NETWORK_MASK must point to the network you want to watch. As an (typical) example, 192.168.1.0/24 would be a value for that. The -o thing tells fing to write to that .csv file (choose a wise name :).

You can listen for different networks by creating different fing instancies. But remember to change the name of the .csvs files, to avoid problems.

The python script listens for every change in the shared folder (shared between all fing instances and the script), but only acts when the change has something to do with .csvs. All information will be stored in the same measure in influxdb (retention policy must be week_rp and measure must be activity) and in the same redis database (with keys prefixed with fingo: and the MAC after). A structure for initial influxdb database can be found in the folder /influxdb.

The simple page has a fully functional dashboard to observe all information stored in the databases. It is optimized enough to sniff around, but maybe not at all for lots of devices in screen. It is a simple node.js app with an index.html that shows the dashboard.

Structure of a deployment:

┌───────┐
│ fing1 │─────────┐
└───────┘         │                                   ┌───────┐
                  │                              ┌────│ redis │─────┐
┌───────┐         │             ┌─────────────┐  │    └───────┘     │ ┌─────────┐
│ fing2 │─────────┼──fing_data──│ fing-parser │──┤                  ├─│ web app │
└───────┘         │   volume    └─────────────┘  │   ┌──────────┐   │ └─────────┘
                  │                              └───│ influxdb │───┘
┌───────┐         │                                  └──────────┘
│ fingN │─────────┘
└───────┘

The docker-compose file

The repository has a docker-compose.yaml file that I'm using right now to deploy the system in my home. Could be a good start for a deployment in your home. For helping to build everything, a bash script called docker.sh is provided that allows you to build, remove, push and pull the images. Maybe the last two are not needed, but well, they are in the script just in case. See ./docker.sh --help to know how to use it, but probably a ./docker.sh build fing fing-parser will be enough.

The monitor service has some tags for a deployment in træfik. Could be useful if you have more web services in your device. Pay attention copying all labels, maybe this configuration is not suitable for you, or directly, don't work.

It is also available a configuration for a chronograf instance. Allows you to inspect the data inside an influxdb database quickly.

Recommendations

As this project is intended to be run in a little computer, like a Raspberry Pi, this recommendations from me to you could be useful:

1. Prepare the images in your computer instead in your destination. To do so:

• You need Docker for macOS (on macOS) or (in Linux) have enabled binfmt and installed qemu for enable ARM code to be executed in you CPU. Some time ago, I wrote about enabling this on Linux in my blog.
• Run ./docker.sh build fing fing-parser.
• Pass your images with a pipe. (uhm?) docker image save fing | ssh pi@raspberry docker image load and docker image save fing-parser | ssh pi@raspberry docker image load to send both images from your computer to the destination computer.

2. Always have a restart policy distinct from no in the docker-compose-yaml. If your device resets for some reason, the containers will restart (instead of keep them stopped).
3. Have fun
4. Maybe I think more of them :)