OpenFeature Ruby SDK

OpenFeature is an open specification that provides a vendor-agnostic, community-driven API for feature flagging that works with your favorite feature flag management tool or in-house solution.

🚀 Quick start

Requirements

Supported Ruby Version	OS
Ruby 3.4.x	Windows, MacOS, Linux
Ruby 4.0.x	Windows, MacOS, Linux

This project supports all Ruby versions in active maintenance per the Ruby maintenance schedule.

Install

Install the gem and add to the application's Gemfile by executing:

bundle add openfeature-sdk

If bundler is not being used to manage dependencies, install the gem by executing:

gem install openfeature-sdk

Usage

require 'open_feature/sdk'

# API Initialization and configuration

OpenFeature::SDK.configure do |config|
  # your provider of choice, which will be used as the default provider
  config.set_provider(OpenFeature::SDK::Provider::InMemoryProvider.new(
    {
      "flag1" => true,
      "flag2" => 1,
      # Variant flags report a variant and reason, and can target on the evaluation context
      "flag3" => OpenFeature::SDK::Provider::InMemoryProvider::Flag.new(
        variants: {"on" => true, "off" => false},
        default_variant: "off",
        context_evaluator: ->(ctx) { "on" if ctx&.targeting_key == "beta-user" }
      )
    }
  ))
end

# Create a client
client = OpenFeature::SDK.build_client

# fetching boolean value feature flag
bool_value = client.fetch_boolean_value(flag_key: 'boolean_flag', default_value: false)

# a details method is also available for more information about the flag evaluation
# see `EvaluationDetails` for more info
bool_details = client.fetch_boolean_details(flag_key: 'boolean_flag', default_value: false)

# fetching string value feature flag
string_value = client.fetch_string_value(flag_key: 'string_flag', default_value: 'default')

# fetching number value feature flag
float_value = client.fetch_number_value(flag_key: 'number_value', default_value: 1.0)
integer_value = client.fetch_number_value(flag_key: 'number_value', default_value: 1)

# get an object value
object = client.fetch_object_value(flag_key: 'object_value', default_value: { name: 'object'})

🌟 Features

Status	Features	Description
✅	Providers	Integrate with a commercial, open source, or in-house feature management tool.
✅	Targeting	Contextually-aware flag evaluation using evaluation context.
✅	Hooks	Add functionality to various stages of the flag evaluation life-cycle.
✅	Logging	Integrate with popular logging packages.
✅	Domains	Logically bind clients with providers.
✅	Eventing	React to state changes in the provider or flag management system.
✅	Shutdown	Gracefully clean up a provider during application shutdown.
✅	Tracking	Associate user actions with feature flag evaluations for experimentation.
✅	Transaction Context Propagation	Set a specific evaluation context for a transaction (e.g. an HTTP request or a thread)
✅	Extending	Extend OpenFeature with custom providers and hooks.

_{Implemented: ✅ | In-progress: ⚠️ | Not implemented yet: ❌}

Providers

Providers are an abstraction between a flag management system and the OpenFeature SDK. Look here for a complete list of available providers.

If the provider you're looking for hasn't been created yet, see the develop a provider section to learn how to build it yourself.

Once you've added a provider as a dependency, it can be registered with OpenFeature like this:

OpenFeature::SDK.configure do |config|
  # your provider of choice, which will be used as the default provider
  config.set_provider(OpenFeature::SDK::Provider::InMemoryProvider.new(
    {
      "v2_enabled" => true,
    }
  ))
end

Blocking Provider Registration

If you need to ensure that a provider is fully initialized before continuing, you can use set_provider_and_wait:

# Using the SDK directly
begin
  OpenFeature::SDK.set_provider_and_wait(my_provider)
  puts "Provider is ready!"
rescue OpenFeature::SDK::ProviderInitializationError => e
  puts "Provider failed to initialize: #{e.message}"
  puts "Error code: #{e.error_code}"
  # original_error contains the underlying exception that caused the initialization failure
  puts "Original error: #{e.original_error}"
end

# Domain-specific provider
OpenFeature::SDK.set_provider_and_wait(my_provider, domain: "feature-flags")

# Via configuration block
OpenFeature::SDK.configure do |config|
  begin
    config.set_provider_and_wait(my_provider)
  rescue OpenFeature::SDK::ProviderInitializationError => e
    # Handle initialization failure
  end
end

The set_provider_and_wait method:

• Waits for the provider's init method to complete successfully
• Raises ProviderInitializationError if initialization fails or times out
• Provides access to the provider instance, error code, and original exception for debugging
• The original_error field contains the underlying exception that caused the initialization failure
• Uses the same thread-safe provider switching as set_provider

In some situations, it may be beneficial to register multiple providers in the same application. This is possible using domains, which is covered in more detail below.

Targeting

Sometimes, the value of a flag must consider some dynamic criteria about the application or user, such as the user's location, IP, email address, or the server's location. In OpenFeature, we refer to this as targeting. If the flag management system you're using supports targeting, you can provide the input data using the evaluation context.

OpenFeature::SDK.configure do |config|
  # you can set a global evaluation context here
  config.evaluation_context = OpenFeature::SDK::EvaluationContext.new("host" => "myhost.com")
end

# Evaluation context can be set on a client as well
client_with_context = OpenFeature::SDK.build_client(
  evaluation_context: OpenFeature::SDK::EvaluationContext.new("controller_name" => "admin")
)

# Invocation evaluation context can also be passed in during flag evaluation.
# During flag evaluation, invocation context takes precedence over client context
# which takes precedence over API (aka global) context.
bool_value = client.fetch_boolean_value(
  flag_key: 'boolean_flag',
  default_value: false,
  evaluation_context: OpenFeature::SDK::EvaluationContext.new("is_friday" => true)
)

Hooks

Hooks allow for custom logic to be added at well-defined points of the flag evaluation life-cycle. Look here for a complete list of available hooks. If the hook you're looking for hasn't been created yet, see the develop a hook section to learn how to build it yourself.

Hooks can be registered at the global, client, or flag invocation level.

# Define a hook
class MyHook
  include OpenFeature::SDK::Hooks::Hook

  def before(hook_context:, hints:)
    puts "Evaluating flag: #{hook_context.flag_key}"
    nil
  end

  def after(hook_context:, evaluation_details:, hints:)
    puts "Flag #{hook_context.flag_key} evaluated to: #{evaluation_details.value}"
  end

  def error(hook_context:, exception:, hints:)
    puts "Error evaluating #{hook_context.flag_key}: #{exception.message}"
  end

  def finally(hook_context:, evaluation_details:, hints:)
    puts "Evaluation complete for #{hook_context.flag_key}"
  end
end

# Register at the API (global) level
OpenFeature::SDK.hooks << MyHook.new

# Register at the client level
client = OpenFeature::SDK.build_client
client.hooks << MyHook.new

# Register at the invocation level
client.fetch_boolean_value(
  flag_key: "my-flag",
  default_value: false,
  hooks: [MyHook.new]
)

Logging

The SDK includes a built-in LoggingHook that provides structured log output for flag evaluations. It logs at the before, after, and error stages of the hook lifecycle.

# Use the SDK's default logger (from Configuration)
OpenFeature::SDK.hooks << OpenFeature::SDK::Hooks::LoggingHook.new

# Or provide your own logger
logger = Logger.new($stdout)
OpenFeature::SDK.hooks << OpenFeature::SDK::Hooks::LoggingHook.new(logger: logger)

# Optionally include evaluation context in log output
OpenFeature::SDK.hooks << OpenFeature::SDK::Hooks::LoggingHook.new(
  logger: logger,
  include_evaluation_context: true
)

Log output uses a structured key=value format:

• before (DEBUG): stage=before domain=my-domain provider_name=my-provider flag_key=my-flag default_value=false
• after (DEBUG): includes reason, variant, and value
• error (ERROR): includes error_code and error_message

Domains

Clients can be assigned to a domain. A domain is a logical identifier which can be used to associate clients with a particular provider. If a domain has no associated provider, the default provider is used. Clients look up their domain's provider on every call, so a client created before set_provider follows the new binding.

OpenFeature::SDK.configure do |config|
  config.set_provider(OpenFeature::SDK::Provider::NoOpProvider.new, domain: "legacy_flags")
end

# Create a client for a different domain, this will use the provider assigned to that domain
legacy_flag_client = OpenFeature::SDK.build_client(domain: "legacy_flags")

Eventing

Events allow you to react to state changes in the provider or underlying flag management system, such as flag definition changes, provider readiness, or error conditions. Providers emit initialization events (PROVIDER_READY on success, PROVIDER_ERROR on failure) from init. A provider without init is READY as soon as it's registered, and its PROVIDER_READY handlers run before set_provider returns. For providers that define init but not emits_lifecycle_events? returning true, the SDK emits PROVIDER_READY after init returns and PROVIDER_ERROR if it raises. That path is deprecated (spec Appendix E), and the SDK logs a deprecation warning once per provider class to its configured logger. Handlers for events a provider emits from init run while init is still in progress, so they can't call set_provider_and_wait for that same provider. Some providers emit additional events: PROVIDER_CONFIGURATION_CHANGED, PROVIDER_STALE, PROVIDER_RECONCILING and PROVIDER_CONTEXT_CHANGED. Handlers receive a hash with the provider_name and any details the provider sent, such as message or error_code.

Please refer to the documentation of the provider you're using to see what events are supported.

# Register event handlers at the API (global) level
ready_handler = ->(event_details) do
  puts "Provider #{event_details[:provider_name]} is ready!"
end

OpenFeature::SDK.add_handler(OpenFeature::SDK::ProviderEvent::PROVIDER_READY, ready_handler)

# Providers emit events using the EventEmitter mixin: lifecycle events from init, and spontaneous
# events to signal internal state changes like configuration updates.
class MyEventAwareProvider
  include OpenFeature::SDK::Provider::EventEmitter

  def emits_lifecycle_events? = true

  def init(evaluation_context)
    # Start background process to monitor for configuration changes
    start_background_process
    emit_event(OpenFeature::SDK::ProviderEvent::PROVIDER_READY)
  end

  def start_background_process
    Thread.new do
      # Monitor for configuration changes and emit events when they occur
      if configuration_changed?
        emit_event(
          OpenFeature::SDK::ProviderEvent::PROVIDER_CONFIGURATION_CHANGED,
          message: "Flag configuration updated"
        )
      end
    end
  end
end

# Remove specific handlers when no longer needed
OpenFeature::SDK.remove_handler(OpenFeature::SDK::ProviderEvent::PROVIDER_READY, ready_handler)

Shutdown

The OpenFeature API provides a shutdown method to perform cleanup of all registered providers. This should only be called when your application is in the process of shutting down.

# Shut down all registered providers and clear state
OpenFeature::SDK.shutdown

The SDK also shuts down a replaced provider once no domain uses it. Events from shut-down providers are ignored.

Individual providers can implement a shutdown method to perform cleanup:

class MyProvider
  def shutdown
    # Perform any shutdown/reclamation steps with flag management system here
  end
end

Tracking

The tracking API allows you to use OpenFeature abstractions and objects to associate user actions with feature flag evaluations. This is essential for robust experimentation powered by feature flags. For example, a flag enhancing the appearance of a UI component might drive user engagement to a new feature; to test this hypothesis, telemetry collected by a hook or provider can be associated with telemetry reported in the client's track function.

client = OpenFeature::SDK.build_client

# Simple tracking event
client.track("checkout_completed")

# With evaluation context
client.track(
  "purchase",
  evaluation_context: OpenFeature::SDK::EvaluationContext.new(targeting_key: "user-123")
)

# With tracking event details (optional numeric value + custom fields)
details = OpenFeature::SDK::TrackingEventDetails.new(
  value: 99.99,
  plan: "premium",
  currency: "USD"
)
client.track("subscription", tracking_event_details: details)

Note that some providers may not support tracking; if the provider does not implement a track method, the call is a no-op. Check the documentation for your provider for more information.

Transaction Context Propagation

Transaction context is a container for transaction-specific evaluation context (e.g. user id, user agent, IP). Transaction context can be set where specific data is available (e.g. an auth service or request handler) and by using the transaction context propagator it will automatically be applied to all flag evaluations within a transaction (e.g. a request or thread).

The SDK ships with a ThreadLocalTransactionContextPropagator that stores context in Thread.current:

# Set up the propagator
OpenFeature::SDK.set_transaction_context_propagator(
  OpenFeature::SDK::ThreadLocalTransactionContextPropagator.new
)

# Set transaction context (e.g. in a request middleware)
OpenFeature::SDK.set_transaction_context(
  OpenFeature::SDK::EvaluationContext.new(targeting_key: "user-123", "email" => "user@example.com")
)

# Transaction context is automatically merged during flag evaluation.
# Merge precedence: invocation > client > transaction > API (global)
client = OpenFeature::SDK.build_client
client.fetch_boolean_value(flag_key: "my-flag", default_value: false)

You can implement a custom propagator by including the TransactionContextPropagator module:

class MyRequestScopedPropagator
  include OpenFeature::SDK::TransactionContextPropagator

  def set_transaction_context(evaluation_context)
    # Store context in your request-scoped storage
    RequestStore[:openfeature_context] = evaluation_context
  end

  def get_transaction_context
    RequestStore[:openfeature_context]
  end
end

Extending

Develop a provider

To develop a provider, you need to create a new project and include the OpenFeature SDK as a dependency. This can be a new repository or included in the existing contrib repository available under the OpenFeature organization. You’ll then need to write the provider by implementing the Provider duck.

class MyProvider
  include OpenFeature::SDK::Provider::EventEmitter

  # Tells the SDK that init emits PROVIDER_READY/PROVIDER_ERROR itself.
  def emits_lifecycle_events? = true

  # Optional: a provider without init is READY as soon as it's registered.
  # Both arguments are optional; declare only the ones you need.
  # `domain` is the domain the provider was bound to, or nil for the default provider.
  def init(evaluation_context = nil, domain: nil)
    # Perform any initialization steps with flag management system here, then emit PROVIDER_READY.
    # Return value is ignored
    # **Note** The OpenFeature spec defines a lifecycle method called `initialize` to be called when a new provider is set.
    # To avoid conflicting with the Ruby `initialize` method, this method should be named `init` when creating a provider.
    emit_event(OpenFeature::SDK::ProviderEvent::PROVIDER_READY)
  rescue => e
    emit_event(OpenFeature::SDK::ProviderEvent::PROVIDER_ERROR, error_code: OpenFeature::SDK::Provider::ErrorCode::GENERAL, message: e.message)
    raise
  end

  # Optional: return true if this provider keeps state for a single domain.
  # The SDK then raises ArgumentError if the same instance is bound to a second domain.
  def domain_scoped?
    false
  end

  def shutdown
    # Perform any shutdown/reclamation steps with flag management system here
    # Return value is ignored
  end

  def fetch_boolean_value(flag_key:, default_value:, evaluation_context: nil)
    # Retrieve a boolean value from provider source
    # To signal an error, return ResolutionDetails with an error_code, or raise
    # OpenFeature::SDK::EvaluationError.new("message", error_code: OpenFeature::SDK::Provider::ErrorCode::FLAG_NOT_FOUND)
  end

  def fetch_string_value(flag_key:, default_value:, evaluation_context: nil)
    # Retrieve a string value from provider source
  end

  def fetch_number_value(flag_key:, default_value:, evaluation_context: nil)
    # Retrieve a numeric value from provider source
  end

  def fetch_integer_value(flag_key:, default_value:, evaluation_context: nil)
    # Retrieve a integer value from provider source
  end

  def fetch_float_value(flag_key:, default_value:, evaluation_context: nil)
    # Retrieve a float value from provider source
  end

  def fetch_object_value(flag_key:, default_value:, evaluation_context: nil)
    # Retrieve a hash value from provider source
  end

  # Optional: implement tracking support (spec 6.1.4)
  # If not defined, Client#track is a no-op
  def track(tracking_event_name, evaluation_context:, tracking_event_details:)
    # Record a tracking event with your flag management system
  end
end

Built a new provider? Let us know so we can add it to the docs!

Develop a hook

To develop a hook, you need to create a new project and include the OpenFeature SDK as a dependency. This can be a new repository or included in the existing contrib repository available under the OpenFeature organization. Implement your own hook by including the OpenFeature::SDK::Hooks::Hook module. You only need to define the stages you care about — unimplemented stages are no-ops by default.

class MyLoggingHook
  include OpenFeature::SDK::Hooks::Hook

  def before(hook_context:, hints:)
    puts "Evaluating #{hook_context.flag_key}"
    nil # Return nil or an EvaluationContext to merge
  end

  def after(hook_context:, evaluation_details:, hints:)
    puts "Result: #{evaluation_details.value}"
  end

  # error and finally are optional — only define what you need
end

Built a new hook? Let us know so we can add it to the docs!

⭐️ Support the project

• Give this repo a ⭐️!
• Follow us on social media:
  • Twitter: @openfeature
  • LinkedIn: OpenFeature
• Join us on Slack
• For more, check out our community page

🤝 Contributing

Interested in contributing? Great, we'd love your help! To get started, take a look at the CONTRIBUTING guide.

Thanks to everyone who has already contributed

Made with contrib.rocks.