Datadog APM: Traces + Metrics

gmarik 1 min

I’ve been working with Datadog’s Application Perfomance Monitoring recently and discovered that the APM agent produces both traces and metrics, which isn’t obvious from looking at the docs.

When the APM is configured, ie for ruby:

require 'ddtrace'
Datadog.configure do |c|
  # This will activate auto-instrumentation for Rails
  c.use :rails

Then the agent starts collecting various rails traces. Also, it collects related metrics (aka trace metrics).

The differenece is important since 1) traces are sampled and metrics aren’t 2) traces provide much more information than metrics

Traces vs Traced Metrics



DD touches on trace metrics but provides no additional information as to where/how it’s available:

Datadog APM computes following aggregate statistics over all the traces instrumented, regardless of sampling:

  • Total requests and requests per second

  • Total errors and errors per second

  • Latency

  • Breakdown of time spent by service/type

Related Posts
Read More
Javascript is my favorite language
Hosting static site with Hugo, Nginx, and Cloud Run
read or add one↓