Get Meaningful Infrastructure Metrics Faster with SignalFx

SignalFx customers love collectd because it serves as a single collection and forwarding point for all the metrics they care about–from the infrastructure metrics of every node to operating metrics from open source platforms like Elasticsearch and custom app metrics through integrations with things like StatsD. While collectd emits these metrics in great detail, users typically want to look at some metrics in aggregate, instead of how they are natively reported. For instance, when CPU time is reported in jiffies and getting to actual utilization requires a computation.

SignalFx - Meaningful Metrics - CPU load average manual computation

Computing load average manually by applying analytics to the native metrics.


For stats like this, to get to the actual desired metric typically means running additional analytics on the native metrics to combine them into something useful for every new visualization, chart, or alert. We’ve now done some of the heavy lifting for you, using SignalFx’s collectd plugin to pre-compute utilization metrics for CPU, memory, disk space, disk I/O, and network traffic.

For example, to compute percent of memory used on a host using the "memory" plugin for collectd, you could use SignalFlow analytics to compute the ratio of "" to the sum of all memory metrics, subtract that number from 1, and multiply by 100 to display as a percent, as follows:

aggregate metrics - browse

This is easy to compute using SignalFlow analytics, but it’s even easier to have the math done for you. Users of the SignalFx collectd plugin (included with every SignalFx install of collectd) can now just choose the metric "memory.utilization", and SignalFx will do the rest.

aggregate metrics - legend

We’ve created the following new metrics from their native counterparts to save you time:

  • cpu.utilization: The percent of CPU used on a host, from the aggregation plugin.
  • memory.utilization: The percent of memory used on this host, from the memory plugin.
  • disk.utilization: The percent of disk space used on this volume, from the df plugin.
  • disk.summary_utilization: The percent of disk space used on all volumes on a host, from the df plugin.
  • The total number of disk read and write operations, summed across all volumes on a host, from the disk plugin.
  • The total number of bytes transmitted and received by all interfaces on a host, from the interface plugin.

aggregate metrics - docs - memory

Check out the new simple utilization metrics and documentation in the Catalog

If you’re already using SignalFx’s collectd plugin, upgrade to the latest version using: yum upgrade signalfx-collectd-plugin (for CentOS and Amazon Linux)apt-get update && apt-get upgrade signalfx-collectd-plugin (for Ubuntu and Debian). 

Start a free trial to begin monitoring your infrastructure with SignalFx today!

Rebecca Tortell
Posted by

Rebecca Tortell

Rebecca is a product manager with many years of experience helping startups make products that users love. Previously she worked at companies like Turn, Playdom, and Disney Interactive.

Get Meaningful Infrastructure Metrics Faster with SignalFx

Show All Tags
Show Less Tags

Join the Discussion