May 17, 2015
Application monitoring doesn't have to be complicated. Introducing
scout_statsd_rack, a drop-in Ruby gem for monitoring key performance metrics in your Rails app.
scout_statsd_rack leverages StatsD for lightweight Rails app monitoring via Rack middleware.
Lets see how fast we can go from no monitoring to a Rails performance dashboard plus alerting.
To make quick work of this, we are going to use Scout, our hosted monitoring service, vs. rolling our own StatsD infrastructure.
The Scout monitoring agent, Scoutd, has a statsd aggregator built in. It will send Scout any StatsD metrics you have defined in your code.
1 Make sure you've got the latest version of Scoutd installed on your application server.
scout_statsd_rack to the applications Gemfile:
3 In your terminal:
$ bundle install
Browse around on your site for a few minutes - and then check your Scout account. Custom metrics will be available to you directly from the header navigation:
...and when creating dashboards:
So what good is a metric if we don't have an alert on it? Let's get notified on the most important app monitoring metric of all: the error rate.
From the Scout UI, navigate to the "All Servers" role click the button to add a trigger template to a Custom Metric:
We're adding this to the special "All Servers" role so that any server that reports this metric will be monitored (whether it exists now or is added later). Pretty cool, huh?
Next, setup the trigger so that it looks like this:
This will trigger an alert when the rate of errors exceeds 5 per-minute.
Do you have a way to trigger exceptions in your app? If so, go back to your app and trigger some more error messages. In a few minutes, you'll get a notification. Wait a couple more minutes, and you'll get a notification that everything has returned back to normal
scout_statsd_rack is a dead simple way to implement lightweight Rails app monitoring. With Scout, you've got a beautiful way to view and customize your metrics - and create robust alerting around them.
With just a few lines of code and couple minutes of work, we've got:
Scout makes StatsD fast & easy!
Thanks for reading, and be sure to follow us on Twitter for more on leveraging StatsD.