Unleash Edge quickstart

This document helps you get started with Unleash Edge locally.

Why Unleash Edge

Unleash Edge is a lightweight layer between your SDKs and your Unleash instance. It exposes the same HTTP interface as the main Unleash API but is built for higher throughput and lower latency.

Performance & UX

Unleash Edge helps you reduce latency for flag resolution by running closer to your users. For example, Unleash Edge could run on a global content-delivery network (CDN) or as part of your cloud or on-premises infrastructure.

Setting up one or more Edge nodes helps you distribute traffic and reduce the load on your Unleash instance. By default, Unleash Edge relies on in-memory caching, but you can configure it to use Redis or the local filesystem.

Security & resilience

From a security standpoint, Unleash Edge lets you expose a single, SDK-compatible endpoint without opening your Unleash instance to the public, thus reducing the attack surface.

At the same time, Unleash Edge provides an additional layer of resilience, so brief upstream hiccups don't disrupt feature delivery.

Optionally, you can set up multiple layers between your application and the Unleash instance (daisy-chaining). This lets you define the optimal configuration for caching and resilience, while adapting Unleash Edge to your architecture's topology.

Offline mode

Unleash Edge can also run in offline mode, so it doesn't need a connection to an upstream Unleash instance. This simplifies local development and helps in environments with limited connectivity.

Prerequisites

Here's what you need before getting started:

1. An Unleash instance running locally or remotely (version 4.15 or later)
2. A valid API token for your Unleash instance
3. Docker installed and running
4. Your preferred Unleash SDK in a sample app

Run Unleash Edge locally

First, make sure your Unleash instance is running (locally or remotely) and generate a new backend API token.

GitHub Release installer

You can install prebuilt binaries for your OS directly from GitHub Releases:

curl --proto '=https' --tlsv1.2 -LsSf https://github.com/Unleash/unleash-edge/releases/download/<version>/unleash-edge-installer.sh | sh

GitHub Releases versioning

Make sure to replace <version> with a valid GitHub Release tag such as v20.1.0.

You can pick a valid version from the list of releases on GitHub.

Then launch it:

unleash-edge edge \
  --strict \
  --upstream-url <your_unleash_instance> \
  --tokens '<your_backend_token>'

Rust toolchain

If you're comfortable with the Rust toolchain, install the CLI with cargo binstall (or from source):

cargo binstall unleash-edge

Then launch it:

unleash-edge edge \
  --strict \
  --upstream-url <your_unleash_instance> \
  --tokens '<your_backend_token>'

Docker

Launch Unleash Edge locally with Docker:

docker run -it \
  -p 3063:3063 \
  -e STRICT=true \
  -e UPSTREAM_URL=<your_unleash_instance> \
  -e TOKENS='<your_backend_token>' \
  unleashorg/unleash-edge \
  edge

Docker Image versioning

By default, the command above uses the latest unleashorg/unleash-edge tag.

You can also pick a specific version from the list of tags on Docker Hub.

Docker Compose

Launch the examples/docker-compose.yml file:

git clone https://github.com/unleash/unleash-edge/
cd unleash-edge/examples
docker compose up

Docker Compose services

Please note that the sample Docker Compose file includes a Redis service for caching. In case you want to run Unleash Edge without Redis, make sure to remove the REDIS_URL environment variable and the Redis service.

Required parameters

Let's break down the parameters you need to replace in the launch command.

<your_unleash_instance>

GitHub Release installer

This is the URL of your Unleash instance.

Use the base URL, e.g. https://app.unleash-hosted.com/testclient or http://localhost:4242.

Rust toolchain

This is the URL of your Unleash instance.

Use the base URL, e.g. https://app.unleash-hosted.com/testclient or http://localhost:4242.

Docker

This is the URL of your Unleash instance.

Use the base URL, e.g. https://app.unleash-hosted.com/testclient.

Important note about Docker and localhost

When using Docker with a local Unleash instance, localhost refers to the container itself, so you cannot reference your Unleash instance with http://localhost:4242. You'll need to use a different hostname, depending on where the Unleash instance is running.

Solution 1: use host.docker.internal

The easiest solution when using Docker Desktop is to reference the host as host.docker.internal.

On Linux, you also need to define the host with --add-host=host.docker.internal:host-gateway, or just use --network=host.

Then you can set your upstream URL to http://host.docker.internal:4242.

Solution 2: use a custom network

If Unleash runs locally on Docker as well, you could run both containers on the same network.

When launching Unleash with docker compose up, you could use the default Compose network named unleash_default and reference the instance as web.

Your launch command becomes --network unleash_default -e UPSTREAM_URL=http://web:4242.

If you prefer decoupling the network and service names, create a custom network:

# define custom network
docker network create unleash-net

# launch Unleash
docker run -d --name unleash --network unleash-net ....

# launch Unleash Edge
docker run -it --network unleash-net -e UPSTREAM_URL=http://unleash:4242 ....

Docker Compose

This parameter is managed for you in the Docker Compose file.

Both containers run on the same network, therefore the instance is referenced using the service name: http://unleash:4242.

---

<your_backend_token>

GitHub Release installer

This API token is required in strict mode, which is recommended since v19.2.

You can generate a new API token by visiting your Unleash instance, under Admin settings > Access control > API access. Click New API token, give it a name, and confirm the default values.

Note: make sure you keep the single quotes in --tokens '...' so the * isn't expanded by your shell.

Rust toolchain

This API token is required in strict mode, which is recommended since v19.2.

You can generate a new API token by visiting your Unleash instance, under Admin settings > Access control > API access. Click New API token, give it a name, and confirm the default values.

Note: make sure you keep the single quotes in --tokens '...' so the * isn't expanded by your shell.

Docker

This API token is required in strict mode, which is recommended since v19.2.

You can generate a new API token by visiting your Unleash instance, under Admin settings > Access control > API access. Click New API token, give it a name, and confirm the default values.

Note: make sure you keep the single quotes in -e TOKENS='...' so the * isn't expanded by your shell.

Docker Compose

This parameter is managed for you in the Docker Compose file.

You can customize the default tokens by editing the Unleash environment variables:

services:
  unleash:
    environment:
      INIT_FRONTEND_API_TOKENS: "default:development.unleash-insecure-frontend-api-token"
      INIT_CLIENT_API_TOKENS: "default:development.unleash-insecure-api-token"

Unleash Edge is automatically configured with the same backend token.

Update your application code

Before you continue, make sure Unleash Edge is running on port 3063.

You can verify this by fetching http://localhost:3063/internal-backstage/health. This endpoint should respond with {"status":"OK"}.

Then you can start updating your application:

1. Identify the code or configuration file where the Unleash instance URL is defined.
2. Update the SDK configuration to use http://localhost:3063.
3. Make sure the SDK configuration includes a valid API token.
4. Restart your app and test it.

.NET

For example, if you're using the .NET example, update the API URL and token in your .env file:

.env

UNLEASH_API_URL=http://localhost:3063/api
UNLEASH_API_TOKEN=<your_backend_token>

Android

For example, if you're using the Android example:

app/src/main/java/com/example/myapplication/MyApplication.kt

  val instance = DefaultUnleash(
      androidContext = this,
      unleashConfig = UnleashConfig.newBuilder(appName = "codesandbox-android")
          .proxyUrl("http://10.0.2.2:3063/api/frontend/")
          .clientKey("<your_frontend_token>")
          .build()
  )

Important note about localhost and device emulators

Please note that you need to use the 10.0.2.2 IP address from the Android emulator to access localhost.

Flutter

For example, if you're using the Flutter example:

unleash_example/lib/main.dart

  var unleash = UnleashClient(
    url: Uri.parse('http://localhost:3063/api/frontend/'),
    clientKey: '<your_frontend_token>',
    refreshInterval: 5,
    appName: 'local-flutter',
  );

Important note about localhost and device emulators

Please note that you need to use the 10.0.2.2 IP address from the Android emulator to access localhost.

You can use localhost from the iOS emulator without issues.

When using a real iOS device, you need to use your computer's IP address. Run ipconfig on Windows or ifconfig on Linux/macOS to identify your IP address.

Go

For example, if you're using the Go example, update the API URL and token in your .env file:

.env

UNLEASH_API_URL=http://localhost:3063/api
UNLEASH_API_TOKEN=<your_backend_token>

Java

For example, if you're using the Java example:

src/main/java/Main.java

  var flag = "example-flag";
  var url = "http://localhost:3063/api/";
  var token = "<your_backend_token>";

JavaScript

For example, if you're using the JavaScript example:

src/index.js

const unleash = new UnleashClient({
    url: "http://localhost:3063/api/frontend/",
    clientKey: "<your_frontend_token>",
    appName: "javascript-codesandbox",
});

Next.js

For example, if you're using the Next.js example, update the API URL and token in your .env file:

.env

NEXT_PUBLIC_UNLEASH_SERVER_API_URL=http://localhost:3063/api
NEXT_PUBLIC_UNLEASH_FRONTEND_API_TOKEN=<your_frontend_token>

Node.js

For example, if you're using the Node.js example, update the API URL and token in your .env file:

.env

UNLEASH_API_URL=http://localhost:3063/api
UNLEASH_API_TOKEN=<your_backend_token>

PHP

For example, if you're using the PHP example, update the API URL and token in your .env file:

.env

UNLEASH_API_URL=http://localhost:3063/api
UNLEASH_API_TOKEN=<your_backend_token>

Python

For example, if you're using the Python example, update the API URL and token in your .env file:

.env

UNLEASH_API_URL=http://localhost:3063/api
UNLEASH_API_TOKEN=<your_backend_token>

React

For example, if you're using the React example:

src/index.tsx

<FlagProvider
  config={{
    url: "http://localhost:3063/api/frontend/", // Unleash Edge running locally
    clientKey: "<your_frontend_token>",
    refreshInterval: 15,
    appName: "codesandbox-react",
  }}
>

Ruby

For example, if you're using the Ruby example, update the API URL and token in your .env file:

.env

UNLEASH_API_URL=http://localhost:3063/api
UNLEASH_API_TOKEN=<your_backend_token>

Rust

For example, if you're using the Rust example, update the API URL and token in your .env file:

.env

UNLEASH_API_URL=http://localhost:3063/api
UNLEASH_API_TOKEN=<your_backend_token>

Svelte

For example, if you're using the Svelte example:

src/routes/+layout.svelte

const config = {
  url: 'http://localhost:3063/api/frontend/',
  clientKey: '<your_frontend_token>',
  refreshInterval: 5,
  metricsInterval: 5,
  appName: 'codesandbox-svelte'
};

Swift

For example, if you're using the Swift example:

Sources/UnleashExample/main.swift

var unleashClient =  UnleashProxyClientSwift.UnleashClientBase(
    unleashUrl: "http://localhost:3063/api/frontend/",
    clientKey: "<your_frontend_token>",
    refreshInterval: 15,
    appName: "codesandbox-swift",
    context: [:]
)

Important note about localhost and device emulators

Please note that you can use localhost from the iOS emulator without issues.

When using a real iOS device, you need to use your computer's IP address. Run ipconfig on Windows or ifconfig on Linux/macOS to identify your IP address.

Vue

For example, if you're using the Vue example:

src/App.vue

const config = {
  url: 'http://localhost:3063/api/frontend/',
  clientKey: '<your_frontend_token>',
  refreshInterval: 5,
  metricsInterval: 5,
  appName: 'codesandbox-vue'
}

Verify your setup

If you run into issues while connecting your SDK to Unleash Edge, the following commands can help you identify the problem.

# is the local Unleash instance running correctly?
curl http://localhost:4242/health

# is the local Unleash Edge instance running correctly?
curl http://localhost:3063/internal-backstage/health
# or via CLI
unleash-edge health

# is data going through? is my token valid?
curl -H "Authorization: <your_backend_token>" http://localhost:3063/api/client/features

You might encounter some of these common issues:

• If Unleash Edge logs show "connection refused" to 127.0.0.1:4242 within Docker, you're pointing at localhost inside the container. Use host.docker.internal or a shared Docker network instead.
• If Unleash Edge logs show "Edge was not able to validate any of the tokens configured at startup" make sure you're using a valid backend token in your startup command.
• If your SDK logs show "401/invalid token" ensure you're using a valid token from your Unleash instance that matches the environment and project you expect.

Next steps

Congratulations, you've successfully set up Unleash Edge locally!

Unleash Edge offers a lot of flexibility and advanced configuration options worth exploring:

1. Offline mode - Learn how to configure Unleash Edge to work without an Unleash instance, using a local features file.
2. Pretrusted tokens - Learn how to explicitly authorize known frontend tokens without upstream validation.
3. Security considerations in production - Learn how to run Unleash Edge in production with best practices for CORS, health checks, and sensitive endpoints.
4. Persistent cache storage - Learn how to enable persistent cache storage with options such as --backup-folder and --redis-url.
5. Advanced CLI configuration - Learn how to customize the CLI behavior with options such as --base-path, --workers, --allow-list, --edge-request-timeout, and --edge-auth-header.