Dropwizard Metrics for your JRuby apps.
Metrica allows you to collect runtime metrics from your application in your production environment. It's mainly designed to work with Ruby on Rails, but it should work well in all kinds of Ruby apps.
Add this line to your application's Gemfile:
gem 'metrica', github: 'sluukkonen/metrica'
And then execute:
If you're using Rails, Metrica comes out-of-the-box with a Rack middleware that collects useful metrics about the performance of your application.
|activeRequests||Counter||The number of requests that are currently being processed.|
|request.total||Timer||The total time spent processing each request.|
|request.controller||Timer||The time spent in the controller (includes time spent rendering the views).|
|request.rendering||Timer||The time spent rendering views.|
|request.db||Timer||The time spent performing SQL queries.|
|request.redis||Timer||The time spent performing Redis commands.|
|statusCodes.ok||Meter||The number of 200 OK responses.|
|statusCodes.created||Meter||The number of 201 Created responses.|
|statusCodes.noContent||Meter||The number of 204 No Content responses.|
|statusCodes.badRequest||Meter||The number of 400 Bad Request responses.|
|statusCodes.notFound||Meter||The number of 404 Not Found responses.|
|statusCodes.error||Meter||The number of 500 Internal Server Error responses.|
To configure Metrica, add an initializer.
# config/initializers/metrica.rb . do |config| config.environments = %w(staging production) # The environments where Metrica should run. config.reporters << ::::. # Publish the metrics to JMX. end
If you're not using Rails, you'll have to call
Metrica.start manually to start
Metrica metric objects are raw Java objects from the Dropwizard Metrics library, so if you are already familiar with it, using metrica should be easy.
See Dropwizard Metrics's Javadoc for a full list of methods in each metric.
A meter measures the rate of events over time (e.g., “requests per second”) as well as 1-, 5-, and 15-minute moving averages.
class RequestProcessor def initialize @meter = .("requests") end def process(request) @meter.mark # Process the request… end end
A counter is just a number you can increment or decrement atomically.
class RequestProcessor def initialize @counter = .("active-requests") end def process(request) @counter.inc # Process the request… ensure @counter.dec end end
A histogram measures the statistical distribution of values in a stream of data. In addition to minimum, maximum, mean, etc., it also measures 50th, 75th, 90th, 95th, 98th, 99th, and 99.9th percentiles.
class RequestProcessor def initialize @histogram = .("request-sizes") end def process(request) @histogram.update(request.size) # Process the request… end end
A Timer measures the rate that a particular piece of code is called and the distribution of its duration.
class RequestProcessor def initialize @timer = Metrica.timer("requests") end def process(request) context = @timer.time # Process the request… ensure context.stop end end end
Metrica also adds a handy
measure method to the
Timer class for ease-of-use.
class RequestProcessor def initialize @timer = .("requests") end def process(request) @timer.measure do # Process the request… end end end
If you want to customize a metric (see Dropwizard Metrics documentation
for more information), you can register a custom metric by using
For example, if you want to use a timer with a
instead of the default
ExponentiallyDecayingReservoir, you can register it
class RequestProcessor def initialize reservoir = ::SlidingTimeWindowReservoir.new(5, ::TimeUnit::SECONDS) @timer = .("requests", ::Timer.new(reservoir)) end def process(request) @timer.measure do # Process the request… end end end
Instrumenting a method
Metrica includes a handy module that allows you to add a Timer around a method. Simply
Metrica::MethodInstrumentation module an mark the method you'd like to
class RequestProcessor extend :: def process(request) # Process the request… end instrument_method :process, "metric-name" end
Collecting metrics isn't enough, you'll also need to report them somewhere. Metrica can report the metrics to the following systems:
Reporting to Console
. do |config| config.reporters << ::::. end
Reporting to JMX
. do |config| config.reporters << ::::JMXReporter.new end
Reporting to Graphite
. do |config| config.reporters << ::::.(hostname: "localhost", port: 2003) end
Reporting to OpenTSDB
. do |config| config.reporters << ::::.(url: "http://localhost:4242") end
Feedback, ideas, bug reports and pull requests are welcome on GitHub at https://github.com/sluukkonen/metrica.
The gem is available as open source under the terms of the MIT License.