Monitor usage of a VPS

Question

I have a VPS for a couple of side projects and testing and I would really like to monitor what's going on with it. Right now im using chartbeat to know how many users are at each website and my provider(wiredtree) have a dashboard where i can check out some of the resources almost in real time. But those 2 tools don't allow me(as a developer) to get information from them to create tools or widgets (or even hardware) that use that information to display something.

I started working on a weekend project to have some leds illuminate in different colors depending on the amount of visitors one of my website is getting, so i mad a simple javascript snippet that is pinging a web service that records that information and then I have a script to another web service that tells me the amount of user that are currently in my site. So after reading some of the questions here i decided to ask; Am i doing something unnecessary? is there any simple tool that would allow me to know those(and more) information?

Answer 1

Monitoring is usually a two part solution. First you want a monitoring and alerting product. In the Linux world, Nagios is king. It's a little complicated to setup, mostly because its config file is very verbose. This will tell you if there is a problem with any part of your system by running checks every minute or so and sending you an alert via SMS or email if something breaks.

The second part is a graphing product. Options include cacti, cricket, ganglia or munin. This allows you to see the trends over time of your system. My choice is Munin, which is incredibly easy to set up and write plugins for. Munin would also give you a nice easy way to get a lot of the kinds of information you would want. You could easily connect to the node over the network and fetch the current value for a plugin, say the current number of apache hits/s and use this for your LED project. Any information that isn't there can be added via a very simple script.

Answer 2

A VPS shouldn't be any different from monitoring a real server, other than your resources might be a little bit more constrained.

What you are doing sounds fine to me for a one off sort of thing. If you really don't think you are not going to be doing a lot of monitoring in the future, Nagios is probably overkill.

If you think you will be doing a lot of monitoring, Nagios is nice because it is easy to write your own scripts and use them with Nagios. So if you want to monitor specific things unique to your application, you can extended Nagios to do this quickly.