Collectd: Great Infrastructure Performance Metrics, but Where Do I Begin?

In the past 24 months, interest in collectd has grown by an order of magnitude, and with good reason. Whether you’re a SysAdmin, SRE, IT Operations expert, or DevOps Engineer who wants to get infrastructure performance metrics, collectd is something you should be interested in deploying. Two questions I recently heard:

Why Should I Try collectd?

Collectd is changing infrastructure performance landscape in a few ways:

  • It’s open source – you can acquire it for free from For many environments, the savings vs. performance tools that charge per server adds up quickly. You are also less dependent on software vendors (including Splunk) on how you want to collect and analyze metrics when you gather those metrics with collectd
  • It’s extensible – building frontend and backend plugins is straightforward, and there are over 100 plugins today
  • It’s lightweight – collectd requires modest resources to run, and you can configure collectd to get just the metrics you need, saving on I/O

How Do I Start Using collectd?

The wiki at gives you pointers, but IMHO it’s not overly focused on people just getting into collectd. To help users get started, we built The Beginner's Guide to collectd—a fast 20-minute read that addresses getting collectd, configuring collectd and analyzing collectd data. 

We also provided some reference links for deeper reading and a short discussion on how to analyze collectd data with Splunk. Download The Beginner's Guide to collectd today and start experimenting with collectd.

We want to hear your thoughts on collectd (or other metrics collection approaches you’re using), and on this beginner’s guide. Please post your comments below and join the discussion.  

Happy Splunking,

Posted by