RESTful Metrics Ruby Client
Tracks your app's custom business metrics in your Ruby apps.
For more detailed instructions, check out our Dev Center.
Install
gem install restful_metrics-ruby
Configure
Setting your API Key
The only step required for initialization is setting your API key. Once it's set, it's stored in memory while your Ruby program runs. You set your API key with the following command:
RestfulMetrics::Client.set_credentials('214c7da8edd333abc78712313918ffe5')
You can skip this step if you've installed the RESTful Metrics Heroku Addon.
Disable Flag
The client also has an optional flag that prevents the client from actually sending data points to the server. This allows you to keep your RESTful Metrics tracking code in place even in your test enviornments. You can disable the client with the following (by default the client is enabled):
RestfulMetrics::Client.disabled = true unless ENV == 'production'
We added the optional conditional check in the above example to illustrate how the flag can be set dynamically during your application's launch.
Asynchronous Mode
The client currently has built-in support for Delayed::Job. If you enable the asynchronous flag the client will automatically queue the data point for transmission to the server at a later time. This is highly recommended for applications that are sensitive to latency.
RestfulMetrics::Client.async = true
If you use a worker library other than Delayed::Job, you can wrap all your RESTful Metrics calls. For example, if we were using Resque we might add this to an initializer:
module CompoundMetric
@queue = :metrics
def self.perform(fqdn, name, values, distinct_id = nil)
RestfulMetrics::Client.add_compound_metric(fqdn, name, values, distinct_id)
end
end
module Metric
@queue = :metrics
def self.perform(fqdn, name, value, distinct_id = nil)
RestfulMetrics::Client.add_metric(fqdn, name, value, distinct_id)
end
end
def restful_metrics_add_data_point(metric_type, app, metric, value, distinct_id=nil)
unless RestfulMetrics::Client.disabled?
Resque.enqueue(metric_type, app, metric, value, distinct_id)
end
end
We've now separated the metrics and compound metrics into their own queue called metrics
. To add a compound metric data point, we call:
restful_metrics_add_data_point(CompoundMetric, "myapp.com", "impression", ["apple juice", "orange juice", "soda"], "fe352fe23e823668e23e7")
Make sure that the async
flag is set to off. This flag only applies to the built-in Delayed::Job support.
Sending Data Points
Metrics
So given the following values for the attributes:
Attribute | Value |
---|---|
Application Identifier | "myapp.com" |
Metric Name | "impression" |
Value | 1 |
Distinct User Identifier | "fe352fe23e823668e23e7" |
You would transmit this data point with the following:
RestfulMetrics::Client.add_metric("myapp.com", "impression", 1, "fe352fe23e823668e23e7")
Compound Metrics
So given the following values for the attributes:
Attribute | Value |
---|---|
Application Identifier | "myapp.com" |
Compound Metric Name | "beverage_search" |
Values | ["apple juice", "orange juice", "soda"] |
Distinct User Identifier | "fe352fe23e823668e23e7" |
You would transmit this data point with the following:
RestfulMetrics::Client.add_compound_metric("myapp.com", "impression", ["apple juice", "orange juice", "soda"], "fe352fe23e823668e23e7")
Copyright
Copyright (c) 2011-2012 RESTful Labs LLC. See LICENSE for details.