DISCLAIMER: The author accepts no responsibility or liability for the use of this software. A high frequency of ICMP requests could be seen as DoS attack. Please ensure you have permission to ping the target servers!
In the default configuration a server will be pinged 10 times with 0.2s between each ping every 60s.

This tool requires net-tools (if not using a docker container) and gnuplot installed on the system.

Configure the servers to be pinged and their label in the graph's legend in the servers.env file:
```
<domain or ip>:<label>
0.0.0.0:localhost
my_server.x:my-server
```
Note that gnuplot will interprets some characters (such as _ underscore) in <label> as mathematical notation.
Run ./files/inet_stat.sh or start a docker container which includes the script with docker-compose up
Plot the data with ./plotter/plot.sh

Configuration

Plot

The ./plotter/plot.sh script offers a variety of options to control the output of the graph.

    usage: SERVER:LABEL [...]
           [--file|-f CSV_FILE] [--limit-y|-l Y_MAX] [--time-zone|-t UTC_HOURS]
           [--start-date|-S DATE] [--end-date|-E DATE] [--show-hours|-H HOURS]
           [--align-legend|-a POSITION]

Examples:

./plotter/plot.sh example.local:example 0.0.0.0:localhost --limit-y 100 --show-hours 5 --end-date "2021-03-03 20:11" --time-zone -5

./plotter/plot.sh example.local:example 0.0.0.0:localhost -S 2021-03-02 -E 2021-03-03 -f data.csv

Run ./plotter/plot.sh --help for details.

Ping

The script ./files/inet_stat.sh which pings the target servers and records the data can be configured via environment variables:

# space seperated list of servers (takes precedence over servers.env)
SERVERS="127.0.0.1
         0.0.0.0
         localhost"

# sets the cycle at which servers are pinged `sleep $CYCLE` [default: 60s]
CYCLE=60s

# sets the amount of times servers are pinged on each cycle [default: 10]
PING_COUNT=10

# sets the output file name [default: ping.csv]
OUTPUT_FILE=ping.csv

# sets the directory path of the output file [default: ./data]
OUTPUT_PATH=./data

# sets the servers.env file path [default: ./servers.env]
SERVER_FILE=./servers.env

# changes the Epoch columns format `date $DATE_FORMAT`, e.g. -Iseconds, [default: +%s]
# ONLY change if you want to use the csv file for another purpose
# any change from the default will prevent plotting using ./plotter/plot.sh
DATE_FORMAT=+%s

These can also be set in the docker-compose.yml in the environment section.

Latency Data

Latency data is saved in data/ping.csv. You can add your own data as long as it has the following format (first line has the column names):

Epoch,Server,PacketLoss,Min,Average,Max,Deviation
1613501993,localhost,0%,87.519,89.085,91.4,1.106
1613501995,0.0.0.0,0%,120.358,122.399,126.254,1.712
1613501997,127.0.0.1,0%,113.125,116.473,125.609,3.311

Viewed in table form:

(UNIX) Epoch	Server	PacketLoss	Min	Average	Max	Deviation
1613501993	localhost	0%	87.519	89.085	91.4	1.106
1613501995	0.0.0.0	0%	120.358	122.399	126.254	1.712
1613501997	127.0.0.1	0%	113.125	116.473	125.609	3.311

Roadmap

Record each server in a separate file to be able to have gnuplot draw lines between each point.
Record into a new csv each day and have ./plotter/plot.sh temporarily merge these based on the start and end date to be plotted. This should cut down on plotting time as gnuplot will not have to read large files anymore.
- Add a flag to keep the merged file permanently.
Automated plotting
- Directly in the Docker Container and server via web server
- Move the plotting into a web interface