Introduction

FeatureManagement.Optimizely connects two feature flag systems:

• Microsoft Feature Management provides the IFeatureManager API and the feature flag evaluation pipeline used by your application.
• Optimizely Feature Experimentation is the source of truth for your feature flags, targeting rules, and experiments.

The library exposes three pieces that plug into the Microsoft Feature Management pipeline:

OptimizelyFeatureDefinitionProvider

An IFeatureDefinitionProvider that lists the features from your Optimizely project (via OptimizelyClient.GetOptimizelyConfig()). Every feature found in Optimizely is registered with Microsoft Feature Management and wired up to be evaluated by the OptimizelyFeatureFilter.

OptimizelyFeatureFilter

An IFeatureFilter (alias Optimizely) that evaluates a feature by asking Optimizely for a decision. It creates an OptimizelyUserContext from the current user supplied by your IOptimizelyUserContextAccessor, calls Decide(featureName), and returns whether the feature is enabled for that user. This is what makes Optimizely's audiences, targeting rules, and experiments take effect.

OptimizelyContextualFeatureFilter

An IContextualFeatureFilter<OptimizelyUserContext> (alias Optimizely) that evaluates a feature using the OptimizelyUserContext passed to IsEnabledAsync. When you pass a user context to IsEnabledAsync, it is used instead of the accessor.

IOptimizelyUserContextAccessor

Your application's implementation of IOptimizelyUserContextAccessor supplies the current OptimizelyUserContext (user id and attributes) to the filter so Optimizely can evaluate targeting rules for that user.

Registration order

The definition provider must be registered before AddFeatureManagement(), and the feature filter must be added after it:

builder.Services.AddOptimizelyFeatureDefinitionProvider(options =>
{
    options.SdkKey = "...";
});

builder.Services
    .AddFeatureManagement()
    .AddOptimizelyFeatureFilter();