monitor-log

This script monitors a site and url, optionally checking that it contains a string value.

All output from this command is piped directly to stdout.

Getting Started

Note: this script relies on bash. It cannot be run as sh.

1. Clone or download this repo.
2. Copy the .env.example file to .env and configure. Each configuration option is commented inline.
3. Run the command in a screen session, configure it with systemd, or run with nohup.

Is it working?

You can check if the script is working by sending SIGUSR1 to the process. When the script starts, it politely prints out its PID along with the suggested kill script to send the signal. This will output a handy "It's working" along with a timestamp to stdout (so check whatever log file you're piping to).

What does it run?

The script is designed to run debugging in under 5 seconds. Currently, it executes the following scripts (in order) with optimizations for fast running:

1. diff to compare the expected string to the returned string
2. tcpdump to show the traffic during the curl request
3. traceroute to show the current path to the server
4. ping to show if the host is up

What will I get?

You'll get the output of these commands, along with timestamps of when things went down and when they came back up.

Here's an example file produced by temporarily overriding the DNS of google.com to 127.0.0.1 in my /etc/hosts file.