Monitoring RHEL with New Relic

Latest response

I've recently been working with a customer who is using New Relic (http://www.newrelic.com). They have a universal 'Server' plugin that you can use. But since it has to work with all Operating Systems, it tends to be very basic.

I quickly learned that one of the more interesting features with New Relic is that you can write your own plugin for just about anything. Essentially any data you can form into a JSON data and POST it to their API, you can record. There are some limitations (they only record data that you can graph, so just integers and floats), but it's a great idea, I think. That also means that your monitoring app's network bandwidth usage is nothing more than the size of those posts. So if you have a server and post 1000 characters once per minute, you have ~1K/minute per server of bandwidth usage.

Probably the biggest asset with New Relic is their UI. It's simple, easy to navigate, pretty and extremely effective.

So I created my own plugin for New Relic to pull more RHEL-specific information. I published it with the New Relic website:
http://newrelic.com/plugins/red-hat-global-support-services/154

Feature Highlights:

  • Per-CPU Utilization Graphing
  • Detailed CPU State Times
  • Much, much more!

You can register and get a free-level account that actually is pretty usable. If you install the plugin, you can then install an RPM and have your servers showing up in New Relic in ~ 60 seconds.

My next steps are to add some additional data (slabinfo, etc.), but I'd love it if people wanted to check it out and make it better. The source code is availabe on Github:

https://github.com/jduncan-rva/newRHELic

Attachments

Responses