.NET provider

Installation

We will first install the OpenFeature SDK and the flagd provider.

.NET Cli

dotnet add package OpenFeature.Contrib.Providers.Flagd

Package Manager

NuGet\Install-Package OpenFeature.Contrib.Providers.Flagd

Package Reference

<PackageReference Include="OpenFeature.Contrib.Providers.Flagd" />

Packet cli

paket add OpenFeature.Contrib.Providers.Flagd

Cake

// Install OpenFeature.Contrib.Providers.Flagd as a Cake Addin
#addin nuget:?package=OpenFeature.Contrib.Providers.Flagd

// Install OpenFeature.Contrib.Providers.Flagd as a Cake Tool
#tool nuget:?package=OpenFeature.Contrib.Providers.Flagd

Using the flagd Provider with the OpenFeature SDK

This example assumes that the flagd server is running locally For example, you can start flagd with the following example configuration:

flagd start --uri https://raw.githubusercontent.com/open-feature/flagd/main/config/samples/example_flags.json

When the flagd service is running, you can use the SDK with the flagd Provider as in the following example console application:

using OpenFeature.Contrib.Providers.Flagd;

namespace OpenFeatureTestApp
{
    class Hello {
        static void Main(string[] args) {
            var flagdProvider = new FlagdProvider(new Uri("http://localhost:8013"));

            // Set the flagdProvider as the provider for the OpenFeature SDK
            OpenFeature.Api.Instance.SetProvider(flagdProvider);

            var client = OpenFeature.Api.Instance.GetClient("my-app");

            var val = client.GetBooleanValueAsync("myBoolFlag", false, null);

            // Print the value of the 'myBoolFlag' feature flag
            System.Console.WriteLine(val.Result.ToString());
        }
    }
}

Configuring the FlagdProvider

The URI of the flagd server to which the flagd Provider connects to can either be passed directly to the constructor, or be configured using the following environment variables:

Option name	Environment variable name	Type	Default	Values
host	FLAGD_HOST	string	localhost
port	FLAGD_PORT	number	8013
tls	FLAGD_TLS	boolean	false
tls certPath	FLAGD_SERVER_CERT_PATH	string
unix socket path	FLAGD_SOCKET_PATH	string
Caching	FLAGD_CACHE	string		lru
Maximum cache size	FLAGD_MAX_CACHE_SIZE	number	10
Maximum event stream retries	FLAGD_MAX_EVENT_STREAM_RETRIES	number	3
Resolver type	FLAGD_RESOLVER	string	rpc	rpc, in-process
Source selector	FLAGD_SOURCE_SELECTOR	string

Note that if FLAGD_SOCKET_PATH is set, this value takes precedence, and the other variables (FLAGD_HOST, FLAGD_PORT, FLAGD_TLS, FLAGD_SERVER_CERT_PATH) are disregarded.

Note that if you are on NET462 through NET48 as the target framework for your project, you are required to enable TLS and supply a certificate path as part of your configuration. This is a limitation Microsoft has documented.

If you rely on the environment variables listed above, you can use the empty constructor which then configures the provider accordingly:

var flagdProvider = new FlagdProvider();

Alternatively, if you would like to pass the URI directly, you can initialise it as follows:

// either communicate with Flagd over HTTP ...
var flagdProvider = new FlagdProvider(new Uri("http://localhost:8013"));

// ... or use the unix:// prefix if the provider should communicate via a unix socket
var unixFlagdProvider = new FlagdProvider(new Uri("unix://socket.tmp"));  

In-process resolver type

The flagd provider also supports the in-process provider mode, which is activated by setting the FLAGD_RESOLVER env var to IN_PROCESS. In this mode, the provider will connect to a service implementing the flagd.sync.v1 interface and subscribe to a feature flag configuration determined by the FLAGD_SOURCE_SELECTOR. After an initial retrieval of the desired flag configuration, the in-process provider will keep the latest known state in memory, meaning that no requests need to be sent over the network for resolving flags that are part of the flag configuration. Updates to the flag configuration will be sent via the grpc event stream established between the in-process provider and the service implementing the flagd.sync.v1 interface (e.g. flagd-proxy).

Example of using the in-process provider mode:

using OpenFeature.Contrib.Providers.Flagd;

namespace OpenFeatureTestApp
{
    class Hello {
        static void Main(string[] args) {

            var flagdConfig = new FlagdConfigBuilder()
                // set the host and port for flagd-proxy
                .WithHost("localhost")
                .WithPort("8015")
                // set the resolver type to 'IN_PROCESS'
                .WithResolverType(ResolverType.IN_PROCESS)
                // provide the flag source selector, e.g. the name of a Flags custom resource which is watched by the flagd-proxy
                .WithSourceSelector("core.openfeature.dev/flags/sample-flags")
                .Build();

            var flagdProvider = new FlagdProvider(flagdConfig);

            // Set the flagdProvider as the provider for the OpenFeature SDK
            OpenFeature.Api.Instance.SetProvider(flagdProvider);

            var client = OpenFeature.Api.Instance.GetClient("my-app");

            var val = client.GetBooleanValueAsync("myBoolFlag", false, null);

            // Print the value of the 'myBoolFlag' feature flag
            System.Console.WriteLine(val.Result.ToString());
        }
    }
}