What's Up With These Monitoring Example Pages? BACK

Quick Story

I have been a huge fan of MRTG since around 1995 or so - wow I'm getting old!

So, I always tell people (professionally and personally) that it's a great strip-graph monitor tool. I've never used it for availability monitoring but some people do.

I got tired of showing off a dataset I collected in 2002 so I finally monitored some stuff at home and (after sanitizing it) sent it up here a few times a day.

In a normal installation the MRTG process runs every 5 minutes and updates the HTML files and PNG graphics. Since I didn't want to bang on my hosted website here I only send up snapshots every 4 hours or so. The sanitized data is created at snapshot time. The full data is hosted on an internal machine with all the gory details (IP addresses, models of systems, etc.).

MRTG makes the files that you see under the Firewall section for the Inside and Outside Interface. The opening page with the date and the clickable image map is hand-made by me using a few Linux tools. You don't need all that, you really only need a front-page of basic HTML to point people at the underlying pages if you want to make it viewable by people.

The interface monitors are the default for MRTG with just a few lines removed to protect the innocent.

The CPU and Firewall Connection monitors are there because some people worry about those metrics - alot.

The TCP monitor is from a shell script that runs on a Linux box that counts the total number of TCP connections and then counts the TCP connections on a particular port that I'm interested in monitoring.

I've started pulling the "current" number on the Host TCP report and changing some of the colors on the graphic used on the front page based on the number of streams. Anytime I have more than 30 "special" streams up it will turn the bottom of the graphic yellow. This is done with ImageMagick. That's really about it.

BACK