Learn how to configure Trifle::Traces for your Ruby application.

Configuration

You don't need to use it with Rails, but you still need to run Trifle::Traces.configure.

Global configuration

If you're running it with Rails, create config/initializers/trifle.rb and configure the gem.

Middleware automatically creates an instance of a tracer for you. To do that automatically, you need to provide tracer_class in your configuration. It defaults to Trifle::Traces::Tracer::Hash, but you can override it with your own custom tracer. You may want to configure custom serializer_class that handles block return values serialization. You can use one of pre-defined serializer classes or define your own. It defaults to Trifle::Traces::Serializer::Inspect serializer.

Trifle::Traces.configure do |config|
  config.tracer_class = Trifle::Traces::Tracer::Hash
  config.serializer_class = Trifle::Traces::Serializer::Json
end

To persist your trace, you need to implement callback(s). Please read more about Callbacks.

Custom configuration

Sometimes you may wanna run tracer outside of a middleware. When doing that, setting up tracer_class is not useful. Instead custom configuration can be passed as a keyword argument to any Tracer directly. This configuration will be then used during persistance callbacks.

config = Trifle::Traces::Configuration.new
config.on(:wrapup) do |tracer|
  entry = MyModel.find_by(id: tracer.reference)
  next if entry.nil?

  entry.update(
    tracer_data: tracer.data,
    state: tracer.state
  )
end

Then you can create an instance of your Tracer with custom configuration.

entry = MyModel.create
tracer = Trifle::Traces::Tracer::Hash.new(key: 'my_key', reference: entry.id, config: config)

tracer.trace('My message')
tracer.wrapup

While you don't need to use callbacks to update your record/entry, I would encourage you to do so. With callbacks you can get live updates with minimal configuration.